[somebee]

I will try to add v2 to that benchmark relatively soon - it is a lot faster than v1 and has a much lower memory footprint and faster initial load time :) Should probably combine it with a few pull requests to the benchmark itself - as I consider it pretty flawed (see another comment here).