Expand end-to-end performance pipeline

We should expand the end-to-end performance pipeline to give us a more complete view of performance-related changes. We will vary GPU type, vertical resolution, and horizontal resolution.

The issue does not name files, tests, or an entry point. Start by locating the existing end-to-end performance pipeline and reading how it currently selects hardware and resolution; done should include comparable runs across GPU types, vertical resolutions, and horizontal resolutions with the expanded results captured.

Written by the indexing model from the issue text.