Is it possible to run open3d GPU rendering via ssh?

[rpartsey]

Checklist

• I have searched for similar issues.
• For Python issues, I have tested with the latest development wheel.
• I have checked the release documentation and the latest documentation (for main branch).

My Question

Open3D is a fantastic tool. I use it as part of the AGILE3D interactive GUI tool for iterative point cloud segmentation visualization.

I'd like to run the open3d-based GUI app on my remote machine with Nvidia GPU (and attached display) and interact with the tool by using my local machine's browser (via WebRTC; o3d.visualization.webrtc_server.enable_webrtc() in my app's code). The long-term goal is to develop an annotation application that could be run on the server and accessed by annotators via their browsers.

I see very slow rendering in a browser when I launch my app (with WebRTC enabled) from the ssh terminal (ssh user@server). However, when I use TeamViewer and run the app via remote control and then access it via my browser, the rendering is smooth and fast. I also noticed that in the first case, GPU is not used at all, but in the second case, I see the GPU utilisation. Also, different OpenGL version is used in both cases: OpenGL(4.5) - when launched from the terminal and OpenGL(4.5) - when launched via remote control using TeamViewer. Could you please explain why this is happening? Is it possible to run open3d GPU rendering via ssh?

Any suggestions are much appreciated. Thank you in advance.