Talk:Sieve of Eratosthenes: Difference between revisions
→Trial division sieves should be moved from here
Line 188:
: * [[Extensible_prime_generator]]: automatically adjust size. The current task doesn't indicate how this is done, and a cursory glance shows examples using trial division, wheel trial division, M-R tests, full SoE reseives, expand by segment SoE, segment SoE iterators, segment SoA iterators, etc.
: I'd like to see a segmented sieve task (generate primes between ''a'' and ''b'' without generating all primes to ''b''), but that's a completely different topic. [[User:Danaj|Danaj]] ([[User talk:Danaj|talk]]) 08:51, 11 September 2014 (UTC)
::Just to correct one thing: "generate primes between ''a'' and ''b'' without generating all primes to ''b''" (to ''a'', you mean?) is '''not''' a segmented sieve. Segmented sieve proceeds by segments, one after another, storing the found primes at a side storage. With just one fixed segment, I've seen it called an "offset" sieve. It is, also, a subtask of "produce all primes above ''a'' indefinitely". If we split the bounded sieves (such that work up to a limit) and the unbounded ones, the two would naturally go each into its respective parent. But even bounded sieve may validly be implemented as segmented a.o.t. monolithic sieve, so perhaps there's no place for this split (and it ''is'' a separate split, monolithic vs segmented). Having them all in one place makes for one compelling story.
::Whatever is decided though, moving the trial division sieves to the trial division pag, first, should be a priority. -- [[User:WillNess|WillNess]] ([[User talk:WillNess|talk]]) 10:07, 11 September 2014 (UTC)
| |||