Skip to content

NVIDIA RTX Virtual Workstation: G2 machine family (L4 GPUs), Ubuntu 22.04

This NVIDIA RTX Virtual Workstation is based on the Google Cloud Compute Engine G2 machine family. A VM based on the G2 machine family includes one to eigth NVIDIA L4 GPUs.

G2 machine are not officially supported by NVIDIA for NVIDIA Omniverse development

To deploy a NVIDIA RTX Virtual Workstation on Google Cloud, you do the following:

  1. Deploy an NVIDIA RTX Virtual Workstation (Ubuntu 22.04) in your project from the Google Cloud Marketplace.

  2. Connect to the NVIDIA RTX Virtual Workstation using SSH. For more information about connecting to a Compute Engine virtual machine (VM), see Connect to Linux VMs.

  3. Clone this repository:

    git clone https://github.com/GoogleCloudPlatform/cloud-solutions && \
cd cloud-solutions

  4. Run the setup-workstation.sh script on the NVIDIA RTX Virtual Workstation:

    projects/nvidia-omniverse-on-google-cloud/g2-development-workstation/ubuntu-22.04/setup-workstation.sh/setup-workstation.sh

    The setup-workstation.sh script does the following to the workstation for NVIDIA Omniverse application development:

    • Verify that NVIDIA drivers are installed
    • Install dependencies
    • Install Docker
    • Install NVIDIA Container Toolkit
    • Install the desktop environment

Connect to the NVIDIA RTX Virtual Workstation remote desktop

To connect to the NVIDIA RTX Virtual Workstation remote desktop, follow the instructions in Create a virtual Windows workstation starting from the Install HP Anyware software section. Ignore the steps prior to the Install HP Anyware software section because they guide you in creating a VM, and you already created one.

Note: If you use Chrome Remote Desktop (CRD) to connect to the development workstation remote desktop, you might experience suboptimal performance because CRD doesn't support hardware acceleration.

What's next