For a relatively novel data model with potentially unusual performance characteristics, the chance of finding a properly done, fair, and generally understandable benchmark is virtually impossible.