[Bug tree-optimization/98526] [11 Regression] Double-counting of reduction cost

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98526

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization
             Status|UNCONFIRMED                 |NEW
   Target Milestone|---                         |11.0
   Last reconfirmed|                            |2021-01-05
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Yeah, the body cost handling should be removed from vect_model_reduction_cost
and instead handled only for the cases using vect_transform_reduction

  /* All but single defuse-cycle optimized, lane-reducing and fold-left
     reductions go through their own vectorizable_* routines.  */
  if (!single_defuse_cycle
      && code != DOT_PROD_EXPR
      && code != WIDEN_SUM_EXPR
      && code != SAD_EXPR
      && reduction_type != FOLD_LEFT_REDUCTION)
    {
...