[Bug tree-optimization/92335] [10/11/12 Regression] sinking of loads happen too early which causes vectorization not to be done

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92335

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization
      Known to work|                            |7.3.0
            Summary|missed transformation to    |[10/11/12 Regression]
                   |branchless                  |sinking of loads happen too
                   |                            |early which causes
                   |                            |vectorization not to be
                   |                            |done
      Known to fail|                            |10.1.0
   Target Milestone|---                         |10.4

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I looked into this and find we are sinking the load of y[i] too early.
That is using -fno-tree-sink, we are able to produce the optimized code.
It would also mean it is a regression even and yes the testcase in comment #3
was able to produce vectorized code in GCC 7.3.0.