I was wondering if it was possible to use a VMware Virtual instance of an Ubuntu , provided the VMware machine has high end RTX GPU’s, to deploy compiled Isaac programs to the Kaya Robot.


Could you describe a bit about what you’d like to do? Avoid having to build a Kaya robot, I’m assuming.

I would say this is probably not a worthwhile approach because a a full Kaya robot is more than the compute node and a Jetson has more custom hardware than an x86/RTX setup. The VM would have to emulate the compute node itself first: aarch64 and Jetson hardware (more than x86 and RTX) all running L4T /JP 4.4.1, and you still wouldn’t have any of the sensor or actuators to talk to.

Hi, to give a bit more detail. I current have a Robotics and Deep Learning Lab, as part of the lab requirements I had about 14 High end Zworkstations ordered with RTX 2080 with Ubuntu 18.04 as the OS. This was Ideal but costly, my lab is now being expanded this year in size. To accommodate this growth, we will either need to order another 12 Zworkstations or I was wondering if we could have a very high end Server with multiple Quadro’s installed and have students use Virtual Ubuntu 18.04 machines via VMware to cross compile Isaac programs and deploy the program remotely to the Kaya robots? Additionally as I also teach Deep Learning having a high end server for teaching this course would be ideal.
Kind regards,

Congratulation on the expansion! What you’re describing is virtual development workstations on a shared machine which seems fine, depending on how you configure VMWare to divvy up the GPUs properly. You might be able to get the virtual environments to run Isaac Sim properly too. Once you have your applications built, you intend for students to deploy them to actual Kaya robots running Jetson. How you do that remote deployment depends on your network configuration and access credentials, but all theoretically possible so far as I understand your goal.