Mrrl / GrailSort

Stable In-place sorting in O(n*log(n)) worst time
MIT License
185 stars 19 forks source link

Even with 512 items buffer it's not fast enough #1

Open Mrrl opened 10 years ago

Mrrl commented 10 years ago

GrailSortWithBuffer() beats std::stable_sort() only at array length of 10 millions or less (at 1 million it's 4/3 times faster, but at 100 millions 10% slower).

Mrrl commented 10 years ago

Fixed some error (with block size selection). Now in the good case (many different keys) sorting is 1.3 times faster even on 100M array.