[Bug target/84482] 437.leslie3d regresses on Haswell and SandyBridge at -O3 and -Ofast with generic march and tuning

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84482

Martin Liška <marxin at gcc dot gnu.org> changed:

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

--- Comment #7 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #6)
> So we vectorize less.  Given the caused revision I wonder if this is the
> same issue as in PR84512 where I just committed a fix for?  So ... fixed?

Yes I can confirm that now it's back fast:

  2538bd2e907580ba(27 Feb 2018 15:25): [took: 52.307s] result: OK
437.leslie3d                                       --       28.0         -- S  

  921e542ba9c6bd82(26 Feb 2018 16:17): [took: 57.477s] result: FAILED (1)
437.leslie3d                                       --       32.9         -- S