Accessing Jetson TX2 remotly to compile and test code


we got two Jetson TX2 for educational purposes.
We would like to access those devices remotly(simultaneously) to test our CUDA codes since not everyone has access to Nvidia graphics by themselves.

Is there any guide out there to set up this kind of environment?

We really appreciate your help.

With kind regards


Hi r3137700, the SSH port (21) is already enabled on the boards, you can login over the network with your user and password. This method is commonplace if you are not sitting at your Jetson and display during development, or have your Jetson deployed on a robot. For convenience, you can mount the Jetson’s filesystem remotely using SSHFS.

One thing you might investigate is adding a virtual desktop. I say this because if you simply use ssh with port forwarding, then it is the host PC which runs the GPU and has the CUDA software/hardware requirements. Basically the X software on the jetson is the interface to the GPU driver, and rendering directly to the host PC without this will cause things to not go as expected.

These URLs might be of interest:

The added advantage is that it won’t matter if the host is Windows, Linux, Mac, etc…if it runs a virtual desktop viewer, then the Jetson will not care about what the remote host is.

Hello. I have a single TX2 host connected to my TP_LINK_POCKET_MR_3020 router operating in 3G/ 4G mode. I need to access my host remotely via ssh (or an alternative) even when I am not on the same network. I tried enabling port forwarding on 22 (port for ssh) on my router, but I still wasn’t able to ping the host’s public IP. Could someone help me configure the router so I am able to accomplish remote access to the host via the internet, please? Thank you.

Duplicate of here:
…check the other thread for replies.