Use Cases:

Multi-Display Output

Display rendered output from Unreal Engine projects on multiple displays.

Here's what you'll need:



On Linux host systems that satisfy the requirements listed at the top of this page, it is possible to run packaged Unreal projects from inside a container with full GPU acceleration and have rendered output displayed on the host system. This is particularly useful when deploying multiple instances of a project to a coordinated set of machines, such as when using the nDisplay system to display output on an LED wall or a CAVE.

nDisplay information coming soon!

This page currently includes only basic information about running packaged Unreal Engine projects in containers and displaying the rendered output on host system displays. Linux support for the Unreal Engine’s nDisplay system is slated for introduction in Unreal Engine 4.27 and the contents of this page will be expanded to include information related to using nDisplay in containers once this new version is released.

Key considerations

Implementation guidelines

Implementation-agnostic commands

Running a container with Docker

To run an interactive container locally with Docker, use the following command (replace unreal-project:latest with the tag for your container image):

# Starts a container running a packaged Unreal project
# (Note that we assume the project is defined as the container's entrypoint)
docker run --rm -ti --gpus=all -v/tmp/.X11-unix:/tmp/.X11-unix:rw -e DISPLAY "unreal-project:latest" # <args>

The individual arguments to the docker run command are explained below:

Your packaged Unreal project should start running and displaying output on the host system’s display. If the project does not start automatically and an interactive bash shell starts instead then you will need to ensure you set the project’s startup script as the container image’s ENTRYPOINT and rebuild the image.

Deploying containers with Kubernetes

When deploying a set of containers with Kubernetes, be sure to include the following in the Pod spec template:

# Set the DISPLAY environment variable to match the host system
    name: DISPLAY
    value: ":0"

# Mount the X11 socket in the appropriate path inside the container
    mountPath: /tmp/.X11-unix
    name: x11-socket

# Source the X11 socket from the appropriate path on the host system
    name: x11-socket
      path: /tmp/.X11-unix

Depending on the configuration of the Kubernetes cluster to which the containers will be deployed, it may also be necessary to specify node selectors or resource requests for GPUs.

Using container images from the ue4-runtime project

The base images from the ue4-runtime project whose tags are suffixed with -virtualgl include the X11 runtime libraries required in order to display rendered output on a host system. Once you have built a container image that extends one of these base images with the files for your packaged Unreal project then you can run containers using the commands from the Implementation-agnostic commands section.

Using container images built from custom Dockerfiles

Check out the Tips for working with the NVIDIA Container Toolkit section of the custom Dockerfile guide for instructions on enabling X11 support in your own container images. Once you have built your container images then you can run them using the commands from the Implementation-agnostic commands section.

This content is landing soon!

Be sure to check out the Unreal Containers Twitter feed for news about content updates.