[Bug rtl-optimization/106200] Shrink-wrapping opportunity releated to function call

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106200

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
      Known to work|                            |13.1.0
   Target Milestone|---                         |13.0
         Resolution|---                         |FIXED

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
In this case, GCC 13 can get rid of the full call to sqrt since GCC knows that
dSquared will be positive (or positive inf or a nan) always .


advance(double, double, double, double):
.LFB1011:
        .cfi_startproc
        fmul    d2, d2, d2
        fmadd   d1, d1, d1, d2
        fmadd   d3, d3, d3, d1
        fsqrt   d1, d3
        fmul    d1, d1, d3
        fdiv    d0, d0, d1
        ret

LLVM 17 can do the same optimization.

A modified testcase:
```

double f(double dSquared, double dt)
{
    double mag = dt / (dSquared * __builtin_sqrt(dSquared));

    return mag;        
}
```

GCC can shrink wrap that starting in GCC 15. so closing as fixed.