How do I efficiently generate a list of primes in Perl 6?

Question

Generating a list of primes is incredibly easy in Perl 6 using is-prime:

my @primes = (^∞).grep: *.is-prime;

This works well enough if you need a relatively small number of primes, but is very inefficient for large numbers, since every number is independently checked.

Is there a way to access Perl 6's built-in prime checking logic to efficiently create a list of primes? Otherwise I'll need to build a sieve myself. Easy enough, but I'm afraid a sieve in high-level Perl 6 code will be almost as inefficient as the code I started with.

Answer 1

If you run your program with --profile, you will see that more than 99% of the time is spent in Int.is-prime. Since that effectively is just a wrapper around nqp::isprime_I(), I have tried to run similar code without the wrapper. But that doesn't change anything noticeably. So the brunt of the work is being done in nqp::isprime_I().

So the only option you really have is to parallelize your searches for primes. In the (nearer) future, hyper would be your friend here. But that is currently at the "initial naive implementation" phase, with a more robust implementation being discussed in: https://gist.github.com/jnthn/6a80a9712fb38b32537f9f0e46fca6d7

Until then, if you want to run things faster, you would have to manually break up the range of values you want to check for primedness and run them within a start block, and collect the results from the resulting Promise.