[Bug tree-optimization/113339] New: `-a/-b` is not simplified to `a/b` if done in seperate statements

[Bug tree-optimization/113339] New: `-a/-b` is not simplified to `a/b` if done in seperate statements

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113339

            Bug ID: 113339
           Summary: `-a/-b` is not simplified to `a/b` if done in seperate
                    statements
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: enhancement
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pinskia at gcc dot gnu.org
  Target Milestone: ---

Take:
```
int unopt(int a, int b) {
   a = -a;
   b = -b;
    return a / b;
}
int opt(int a, int b) {
    return -a / -b;
}
```
I would have expected these 2 would produce the same assembly but only opt is
optimized to a/b and the neg is removed.

Note LLVM does not do either: https://github.com/llvm/llvm-project/issues/77717
.

[Bug tree-optimization/113339] `-a/-b` is not simplified to `a/b` if done in seperate statements

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113339

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |pinskia at gcc dot gnu.org
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2024-01-11

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This is currently done only in fold:
```
      /* (-A) / (-B) -> A / B  */
      if (TREE_CODE (arg0) == NEGATE_EXPR && negate_expr_p (arg1))
        return fold_build2_loc (loc, RDIV_EXPR, type,
                            TREE_OPERAND (arg0, 0),
                            negate_expr (arg1));
      if (TREE_CODE (arg1) == NEGATE_EXPR && negate_expr_p (arg0))
        return fold_build2_loc (loc, RDIV_EXPR, type,
                            negate_expr (arg0),
                            TREE_OPERAND (arg1, 0));
      return NULL_TREE;
```

I am going to handle the simple cases for GCC 15 here ...

[Bug tree-optimization/113339] `-a/-b` is not simplified to `a/b` if done in seperate statements

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113339

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Note the fold was added here:
https://gcc.gnu.org/pipermail/gcc-patches/1999-October/020476.html .

[Bug tree-optimization/113339] `-a/-b` is not simplified to `a/b` if done in seperate statements

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113339

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So I looked into the wrong part of fold, but anyways PR 23669 added the folding
to fold instead (and I just noticed I implemented it originally).