User talk:Elibarzilay: Difference between revisions
→Sieve vs trial division.: wheel is finite
(→Sieve vs trial division.: a reply) |
(→Sieve vs trial division.: wheel is finite) |
||
Line 34:
Hello, I'd also like to participate in this discussion. I've added my reply on [[Talk:Sieve of Eratosthenes#Trial division sieves should be moved from here|the page's talk page]]. Let's continue the discussion there, where it is more likely to be noticed. Thanks. -- [[User:WillNess|WillNess]] ([[User talk:WillNess|talk]]) 20:00, 10 September 2014 (UTC)
BTW, about your remark that Turner's sieve calculates the wheels as it operates. I don't think it's right, because a wheel is finite - its when we ''roll'' a wheel that we get the infinite stream, equally calculated by the Turner sieve at each step. But the wheel itself is finite - [3] for 2; [5,7] for 2*3; [7,11,13,17,19,23, 29,31 ] for 2*3*5, etc. And the point to it is that we ''roll'' it by repeated additions, without any need for repeated test divisions. That's where the savings come from, it wouldn't be an optimization otherwise. -- [[User:WillNess|WillNess]] ([[User talk:WillNess|talk]]) 21:42, 10 September 2014 (UTC)
| |||