Add a divisor (operations per invocation) to the benchmark config

[nothing-wrong-with-linux]

Following #69, add a functionality to let users run more than one operation per benchmark and then just divide the time by the user-provided value. Some of the users know very well what they are doing and want to strip excessive side effects or trying to understand what is the iteration cost if their process ends with some kind of aggregation.

[nothing-wrong-with-linux]

Looking at the hpp file, the divisor is there, but i don't see anything like that in the generated js code.