Anonymous user
Talk:Sieve of Eratosthenes: Difference between revisions
→Trial division sieves should be moved from here
Line 237:
::::Elibarzilay: so I've switched back my postponed version for your postponed Bird's, and placed it ''before'' Bird's. You're welcome to tweak it with your ''when-bigger'', or I'll do it when I have more time (will have to check if it's faster or not). It's more-or-less equivalent to Haskell's <code>span</code>-using versions; but with Haskell's GHC, manually inlining the ''span'' as I did here, was more efficient. -- [[User:WillNess|WillNess]] ([[User talk:WillNess|talk]]) 10:09, 14 September 2014 (UTC)
::::: WillNess: I've fixed some of the indentation problems, and also removed the <code>#lang lazy</code> from partial sources (they're misleading, since the <code>#lang</code> should be used at the top of a module and determines its language, it's not just a typographical annotation of the text...). As for the speed, factors are obviously not interesting in theory, and high factors are interesting in practice only. But I didn't measure them or tried to analyze the runtime, I just tried a few semi-long runs since I'm too lazy to invest more time in this... You're welcome to do more detailed analysis of course (and to mention it: something that I didn't do since I didn't put in the effort...). Oh, and btw, my <code>when-bigger</code> trick is to avoid calling another function that will call itself recursively to collect two values, since this kind of thing would be disastrously slow in Lazy Racket (I should know, I implemented it...). My solution was a haskell-span-like function that takes a callback for the tail, which is enough here, and should generally be fast enough and clear. --[[User:Elibarzilay|Elibarzilay]] ([[User talk:Elibarzilay|talk]]) 05:45, 15 September 2014 (UTC)
| |||