TOOL STRUCTURE
The Auto Perf tool hooks into Unreal Engine’s automation test framework, allowing us to launch it with the click of a button. The test reads from data tables to determine the performance test’s parameters, including:
- Minimum duration to wait for the game to stabilize before sampling performance
- Sampling duration
- Window resolution of test
- Global quality setting presets
- Which maps allow for performance testing
Cameras are hand-placed around levels, ideally showing angles players are likely to encounter in game. Each camera has a tag applied to it, identifying it as relevant to performance testing. When the test executes, we iterate through each camera, taking a screenshot and recording how fast our threads are running.
After all data has been collected, we write the output to a new file containing:
- Build version
- Date of test
- Map name
- Resolution captured
- …and all data from each camera
Automated testing was extremely useful for the Wayfinder team, allowing for easy evaluation of our performance targets. We could place as many cameras as needed, and from referencing the collected data and the screenshots, we could investigate areas experiencing slowdowns.
TOOL STRUCTURE
The Auto Perf tool hooks into Unreal Engine’s automation test framework, allowing us to launch it with the click of a button. The test reads from data tables to determine the performance test’s parameters, including:
- Minimum duration to wait for the game to stabilize before sampling performance
- Sampling duration
- Window resolution of test
- Global quality setting presets
- Which maps allow for performance testing
Cameras are hand-placed around levels, ideally showing angles players are likely to encounter in game. Each camera has a tag applied to it, identifying it as relevant to performance testing. When the test executes, we iterate through each camera, taking a screenshot and recording how fast our threads are running.
After all data has been collected, we write the output to a new file containing:
- Build version
- Date of test
- Map name
- Resolution captured
- …and all data from each camera
Automated testing was extremely useful for the Wayfinder team, allowing for easy evaluation of our performance targets. We could place as many cameras as needed, and from referencing the collected data and the screenshots, we could investigate areas experiencing slowdowns.
