Talk:Chernick's Carmichael numbers: Difference between revisions

</pre>

Remembering that prime numbers cannot end in 5 from the above I deduce that if m is not divisible by 320 then at least 1 of the polynomials must have a non prime result. This reduces the work to be done by 80%. Using this I obtained an answer in 6h, 31min, 16,387 ms. This could easily be halved using pseudo-primes and validating the final result. There is an obvious multi-threading strategy which could cut the time to 1/number of threads. More interestingly programming wise is start prime testing for a given m from the 10th polynomial. Observe that if this is not prime then for m*2 the 9th polynomial can not be prime (it is the same number). Similarly for m*4 the 8th polynomial can not be prime ... down to m*128 for the 3rd polynomial and then m*384 for the first polynomial. If the 10th polynomial is prime but the Nth (down to the 3rd) is not then the same applies just shifted. Keeping track of the m that do not need to be tested should reduce the time significantly.--[[User:Nigel Galloway|Nigel Galloway]] ([[User talk:Nigel Galloway|talk]]) 09:34, 6 June 2019 (UTC)