tinkered with timsort a bit a few months ago. comparing that to your sort, I get numbers like
xinokSort random: 77 ascending: 0 descending: 21 timsort random: 354 ascending: 1 descending: 4 where each are sorting a 500k element array of int, times are msecs, compilation flags were -O -inline -release, sources are http://personal.utulsa.edu/~ellery-newcomer/timsort.d http://personal.utulsa.edu/~ellery-newcomer/xinokSort.d Nice job, Xinok. anyone want to try to optimize my timsort? :)
