[Bug target/84201] 549.fotonik3d_r from SPEC2017 fails verification with recent Intel and AMD CPUs

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84201

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|NEW                         |RESOLVED

--- Comment #19 from Richard Biener <rguenth at gcc dot gnu.org> ---
There's now --param vect-induction-float=0 available as mitigation which
disables all vectorization of floating point typed inductions.

There's nothing wrong with GCC as for the original report.  Though we attempt
to make -Ofast usable for SPEC this particular vectorization isn't off the
hooks and the rounding errors are well within what you'd expect.

So, INVALID for the bug.