Using Docker#
Pulling the image#
The prebuilt image is available on the GitHub container registry. You can pull the image with a tag matching your desired application, or you can build the Docker image locally.
Caution
Only the devel and runtime prebuilt images are directly available from the
GitHub container registry due to the licensing restrictions of the NVIDIA OptiX
library bundled in the other tagged images.
Note that you can still build these images locally from source.
There are four different Docker image variants:
devel, suitable for development, DYNO must be built manually inside the container.runtime, suitable for deployment, DYNO is built and its executable is available in PATH.ros-devel, suitable for development of the ROS node. The node must be built manually inside the container.ros-runtime, suitable for deployment of the ROS node. The DYNO ros node is built and available in the local ROS workspace.
Moreover, all images are available in their default and CUDA (labelled with the
trailing -cuda) variants. The latter are much larger, but add support for
sensors and GPU-accelerated visualization for compatible hosts. Note that ros
images inherit from the cuda image, as they require sensor simulation.
The latest built version is labelled as
-stable.Specific versions are labelled based on to the commits they are build against (
<hash>).
To pull a specific Docker image, use the following command:
docker pull gchr.io/au-agv/dyno:{tag}}
Below is a list of the available tag and a brief description of each.
Note
If you are an end-user looking to run DYNO, the runtime image is suitable for
your application.
Tag |
Description |
|---|---|
|
Development image |
|
Development image, with CUDA support |
|
Runtime image with latest stable version |
|
Runtime image with latest stable version, with CUDA support |
|
ROS interface node image |
|
ROS interface node image, with CUDA support |
Running the image#
You can spawn a new interactive shell with dyno in a container by running:
docker run -it ghcr.io/au-agv/dyno:runtime
You may then use the command dyno to run your scenario. For example, to
run the straight line acceleration scenario with default parameters:
dyno --scenario straightLineAcceleration
Refer to the user guide for more details on the command line interface options and arguments.