This is the code used for the post "Primes, parallelism, and caches Part 2" on my site "Excursions with C++" reachable on https://excursionswithcpp.com. The main branch is matching the post, whereas the alternative branches contains experiments with further optimizations, of which taking over the memory allocation by the program did not give any improvements, but ignoring all the even numbers, which cannot be a prime, gave give improvements. Try it out, copy, change and do experiments. Have fun!