This has been discussed before in bug3661
Have raised a pull request to get this code accepted, it is now up to the library folk to optimise this routine, but I believe it's something that can't be vectorised, so there will probably be no performance benefit.
https://github.com/D-Programming-Language/dmd/pull/325
Regards
Comment #2 by ibuclaw — 2011-08-21T03:28:23Z
*** This issue has been marked as a duplicate of issue 3661 ***