http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54075
--- Comment #21 from François Dumont <fdumont at gcc dot gnu.org> 2012-07-27 07:57:59 UTC --- I haven't touch the grow speed for the moment. I prefer to fix the reserve Standard conformity first. Now I can restore the 4.6 grow speed as it seems to be a relatively correct one. Note that 3x slower with so many more rehash operations doesn't seem so bad ! Of course growing faster will also induce higher memory consumption which is not shown by the simple benchmark proposed here.
