Based on "Can My Water Cooled Raspberry Pi Cluster Beat My MacBook?" (https://youtu.be/hGyJTcdfR1E)
I wanted to test this on my new M1 Mac mini. Then I started to look into Python parallelization and native arm M1 support (there is none for python3 yet). So I wrote it in C++. Also implemented a (dumb) prime number checker to test.
My Mac mini M1 does 100'000 terms for Pi in 850 microseconds when in parallel mode. For all prime numbers up to 1'000'000 it uses 21574 milliseconds.