Running the benchmark

These pages instruct you how to run this yourself on the provided platforms for reproduction and additional platforms for further comparison. See results and platform comparison pages for benchmark evaluations.

What you need

Two machines on a wired network:

• A developer or external machine running the Gazebo warehouse simulation and the robot’s simulated sensors in a Docker container. Keeping the simulation off the machine being measured is what makes the utilization numbers represent the run-time deployed application load.
• The compute platform under test, running the AMR autonomy container, the AI workload container, the simulated sensor-driver load, and the metrics capture script.

DDS is restricted to the wired Ethernet interface so discovery traffic and interference stay out of the measurement.

Where to start

If you want to…	Go to
Reproduce the published results	Run the benchmark
Configure a specific platform	Platform setup
Benchmark new platforms	Adding a platform