How this blew up into such a shitstorm from a single test involving just two devices is beyond me. It's a bit like taking a single person at random from Country A and a single person at random from Country B and then declaring the winner of a single race as proof positive that the entire country is better at running. I realize it's Apple and that means even minor things are fucking magical and revolutionary or complete catastrophes, but this is a new low.
Even if the processors were all from the same company, we would still expect to see variation in performance due to how CPUs are made. Anyone who's done any overclocking knows that there's going to be a noticeable performance difference between CPUs with the same exact model number, simply because of the binning process used by the manufacturers like Intel, TSMC, etc. Maybe the CPU you got barely qualified for a particular bin and you can't get a lot of additional performance out of it or perhaps you're on the other end where the chip you got was barely not good enough for the next bin and you can get a reasonable amount of extra performance out of it.
You'd want to sample dozens of different devices in a controlled setting to get the complete picture and to see how the distributions overlap. It could be that on average TSMC does get a slight advantage over a Samsung chip, but that in this one particular test, we got a Samsung SoC that is poorer than average for a Samsung manufactured chip and a TSMC chip that is better than average for a TSMC manufactured chip.