[PATCH] Add math benchmark latency test

[Alexander Monakov]

On Wed, 16 Aug 2017, Arjan van de Ven wrote:
> > Ah sorry, I should have been clearer - inverse in terms of units, not
> > value, i.e. throughput is typically expressed in items/unit time while
> > latency is expressed in time/unit item.
> 
> maybe "throughput-time" ?

I suggest using "reciprocal throughput" if you're looking for a short term
for 'independent executions per unit time'.  It's easier to recognize and
already used in practice (e.g. in docs by Agner Fog).

Alexander