Since we only get the output at the end of the execution, we can't eyeball the output streaming by to gauge where time is being spent