--- Comment #2 from 2011-07-06 10:42:20 PDT ---
Was this update in the sort code caused by this enhancement request, or are
they unrelated?

Timings with DMD 2.054beta, a different CPU:

sort-sort2 benchmarks (milliseconds), N=6000000:
  Random distribution:
    sort:    1892
    sort2:   1131
  Already sorted arrays:
    sort:     748
    sort2:    376
  Inverted order arrays:
    sort:    1048
    sort2:    656
  Few random doubles appended to the sorted arrays:
    sort:    3085
    sort2:    734

It seems the random case is improved a lot, the already sorted case is
improved, the inverted order is about the same, and the few random values added
case is slower than before.

Configure issuemail:
------- You are receiving this mail because: -------

Reply via email to