tasty-bench for benchmarks

[reubenharry]

Benchmarking performance is pretty important for monad-bayes (for example, I have thrown out attempted refactorings based on worse benchmark performance). However, I found criterion pretty hard to use, and the workflow for comparing to previous bench results was not great (probably in part because of lack of knowledge). I just watched https://www.youtube.com/watch?v=EHVz5j1gYkk&list=PLBqWQH1MiwBTwo2wrwINAorzXRumxO41s&index=19&ab_channel=Konfy and I wonder whether switching to tasty for tests and tasty-bench might be a nice medium term improvement.