Performance scaling / testing

[aram-devdocs]

Proof is in the pudding. The goal here is to build something that is more efficient for use cases where Prisma can be too heavy. Lets set up some testing to ensure that we are the better option for lightweight types, and a way to potentially test user models for their performance score. this data can help us long term auto assign levels of detail, similar to a way video games do a graphics test before setting up your settings for you to AUTO. If we can take in desired traits and balance that against optimal settings, we can validate any trade offs using these tools