I was working with an optimisation problem based around cplex a few years ago that took about 5 minutes to complete - at the time I worked out that if we'd star

I was working with an optimisation problem based around cplex a few years ago that took about 5 minutes to complete - at the time I worked out that if we'd started the optimisation on a machine 10 years prior, it would have been quicker to just wait until the present day (of this story) and then use the code we were writing because improvements in the algorithm and in the hardware added up to a million-fold improvement in performance! If I remember the timelines correctly I think the original version would still have been running today even.