[esterna]

Very nice! One small note: In an Eratosthenes sieve, you can start crossing out numbers from p^2, since i*p for i<p will have already been crossed out in step i at the latest. You can simply replace

  for (size_t i=2; i <= 0xFFFFFFFF / p; i++) {

by

  for (size_t i=p; i <= 0xFFFFFFFF / p; i++) {

[adgjlsfhk1]

and you can get another 2x improvement by incrementing i by 2

for (size_t i=p; i <= 0xFFFFFFFF / p; i+=2) {

since every other one will be a multiple of 2