[Bug tree-optimization/128489] New: [17 Regression] match pattern introduces CONVERT_EXPR instead of VCE for vectors

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128489

            Bug ID: 128489
           Summary: [17 Regression] match pattern introduces CONVERT_EXPR
                    instead of VCE for vectors
           Product: gcc
           Version: 17.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: pinskia at gcc dot gnu.org
          Reporter: pinskia at gcc dot gnu.org
  Target Milestone: ---

Created attachment 66101
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=66101&action=edit
testcase

This match pattern:
```
/* x != CST1 ? x + CST2 : CST3 -> x + CST2 when CST1 + CST2 == CST3.
   This handles both scalars and uniform vectors.  */
(for cnd (cond vec_cond)
 (simplify
  (cnd (ne @0 uniform_integer_cst_p@1)
       (plus@2 @0 uniform_integer_cst_p@3)
       uniform_integer_cst_p@4)
  (with {
    tree cst1 = uniform_integer_cst_p (@1);
    tree cst2 = uniform_integer_cst_p (@3);
    tree cst3 = uniform_integer_cst_p (@4);
    wi::overflow_type ovf;
   }
   (if (wi::add (wi::to_wide (cst1), wi::to_wide (cst2),
                 TYPE_SIGN (TREE_TYPE (cst1)), &ovf) == wi::to_wide (cst3))
    (if (!ovf || TYPE_OVERFLOW_WRAPS (TREE_TYPE (cst1)))
     @2
     (with {
       tree utype = unsigned_type_for (type);
      }
      (convert (plus (convert:utype @0) (convert:utype @3)))))))))

```

Should be using view_convert instead of convert.  Most places expect VCE rather
than convert (which seemily acts like a copy for big vectors)
Testcase attached.