in reply to Re: Speeding permutation counting
in thread Speeding permutation counting

Nevermind. It was an interesting idea, but it sucks.

Rate ikegami albert browseruk ikegami 5.76/s -- -26% -89% albert 7.84/s 36% -- -86% browseruk 54.8/s 851% 599% --

Both ikegami and albert's version will be faster per string when processing 1000 strings instead of 100 (since they both have a big overhead), but not enough to even approach BrowserUk's version.

Benchmark code: