[Bug tree-optimization/118132] Vectorize optimization generates obsolete code, which causes significant performance drop

[pinskia at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118132

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
      Known to work|                            |12.1.0
             Status|UNCONFIRMED                 |RESOLVED
      Known to fail|                            |11.4.0
   Target Milestone|---                         |12.0

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So the options needed are `-O3 -fPIE`.



GCC 12 produces:

_Z16GetDispatchTablev:
.LFB16:
        .cfi_startproc
        movzx   eax, BYTE PTR _ZGVZ16GetDispatchTablevE2_v[rip]
        test    al, al
        je      .L18
        mov     rax, QWORD PTR _ZZ16GetDispatchTablevE2_v[rip]
        ret


Which is perfect.

Fixed by r12-897-gde56f95afaaa22 which adds a sink pass after vectorization .