Build a “tx / s” benchmark

[joshua-spacetime]

George Kulakowski comments:
This shouldn’t be too trivial (ie not inserting a single int into an empty table) but also not quite as big as some of the bitcraft reducers can get. Rather, the goal is to have a plausible number that we could have as like the “sticker price” of the db: we do up to NNN tx / s.

cc @tyler Cloutier

In particular, this also needs to go across the wasm boundary.

Edit pgoldman:
#345 adds tx/s measurement for the benchmarks added by #289 .

[kazimuth]

stdb_module/disk/update_bulk/person/unique/load=1000/count=100

On my machine, we are currently at 1000 tx/s