Talk:Penta-power prime seeds

Just thought I'd mention, the Penta-power prime seeds must be odd as n^0+n+1 = n+2 can only be prime if n is odd if n >= 1. --Tigerofdarkness (talk) 08:33, 20 August 2022 (UTC)

Right... and this is probably most efficiently implemented as a sequence of tests, one for each polynomial -- the smaller polynomials filter out so much that parallelism costs more than it would contribute. The stress here is primality testing on large numbers (the task requires testing 83 bit unsigned integers and the stretch goal requires testing 94 bit unsigned integers).

Most (not all) tasks of this nature put primality testing of integers larger than 2^63 in the stretch goal. --Rdm (talk) 16:55, 20 August 2022 (UTC)

Well, maybe it isn't the biggest optimisation of all time but any optimisation is surely useful - in order to achieve the stretch goal, numbers to over 10 million must be considered - even if the primality test almost instantly rejects the even numbers, that's still around 5 000 000 wasted tests. As it happens, it didn't make that much difference to the Algol 68 sample's run time (around 2%) but it did make a difference. Perhaps you could see what checking odd and even numbers does to your runtime, when you have a solution ? --Tigerofdarkness (talk) 18:12, 20 August 2022 (UTC)