[Bug tree-optimization/83232] fma3d spec2000 regression on zen with -Ofast (generic tuning) after r255268 by missed SLP oppurtunity

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83232

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2017-12-01
            Version|unknown                     |8.0
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot 
gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
I belive group detection is confused by those

  einc[4] = 0.0;
  einc[5] = 0.0;

after the vector code.  Like we don't SLP vectorize

double a[2];
double b[2];
void foo (double *p, double *q)
{
  a[0] = b[0];
  a[1] = b[1];
  *p = *q;
  a[0] = b[1];
  a[1] = b[0];
}

I have an idea how to fix that.  Have to first confirm this is the fma3d issue.