[Bug tree-optimization/63644] Kahan Summation with fast-math, pattern not always recognized

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63644

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
It is kinda of vectorized:

  <bb 5>:
  # vect_vec_iv_.37_29 = PHI <vect_vec_iv_.37_30(5), {
1.00000001490116119384765625e-1, 2.0000000298023223876953125e-1,
3.00000011920928955078125e-1, 4.000000059604644775390625e-1 }(4)>
  # ivtmp.43_16 = PHI <ivtmp.43_13(5), 0(4)>
  vect_vec_iv_.37_30 = vect_vec_iv_.37_29 + { 4.000000059604644775390625e-1,
4.000000059604644775390625e-1, 4.000000059604644775390625e-1,
4.000000059604644775390625e-1 };
  vect_s_10.38_32 = vect_vec_iv_.37_29 + { 1.00000001490116119384765625e-1,
1.00000001490116119384765625e-1, 1.00000001490116119384765625e-1,
1.00000001490116119384765625e-1 };
  ivtmp.43_13 = ivtmp.43_16 + 1;
  if (ivtmp.43_13 >= bnd.33_20)
    goto <bb 7>;
  else
    goto <bb 5>;

But then the vectorization part is removed:
  <bb 7>:
  _33 = BIT_FIELD_REF <vect_s_10.38_32, 32, 96>;


This is odd.