Talk:Prime Sieve of Eratosthenes

I made some changes, but I see two problems: --Aboyner 17:38, 4 Mar 2006 (EST)
 * 1) Not everyone might like to have so much explanation for so simple an algorithm
 * 2) I've made it rather heavy on markup. Although it doesn't look very good to me either, I think that it's a limitation of the browsers; probably some settings will make it look better. At any rate, I think that MathML rendering will eventually improve, maybe the article will look good then :)

I like your "crossing out" data structure. Do you have any timings of your algorithm vs the conventional one? I'd imagine it would end up slowing things down, just because the increased memory usage (2 ints and one bool vs a single packed bool) is a big blowup in memory. --Rrenaud 10:47, 5 Mar 2006 (EST)
 * Yeah, my program was about 20 times slower than the usual sieve, for N=10 million. I even tried making them both do the same initialisation, but it's still 8 times slower. --Aboyner 10:17, 6 Mar 2006 (EST)