http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57511
--- Comment #1 from Alexander Monakov <amonakov at gcc dot gnu.org> --- The loop invokes signed integer overflow, but changing 10000 to 10 still keeps the missed optimization there.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57511
--- Comment #1 from Alexander Monakov <amonakov at gcc dot gnu.org> --- The loop invokes signed integer overflow, but changing 10000 to 10 still keeps the missed optimization there.