[Bug tree-optimization/46085] [4..6 Regression] gcc.dg/vect/fast-math-vect-reduc-[57].c failed with -mavx -ffast-math -O3

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46085

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2010-10-19 19:48:35 
UTC ---
[...@gnu-18 gcc]$ cat x.c
extern void abort ();

float b[16] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45};
float c[16] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};

int main ()
{
  int i;
  float diff = 2;

  for (i = 0; i < 16; i++) {
    diff += (b[i] - c[i]);
  }

  if (diff != 242)
    abort ();

  return 0;
}
[...@gnu-18 gcc]$  ./xgcc -B./ x.c -ftree-vectorize -g -mavx -O  -ffast-math 
[...@gnu-18 gcc]$ ./a.out 
Aborted
[...@gnu-18 gcc]$  ./xgcc -B./ x.c -ftree-vectorize -g -mavx -O -ffast-math    
-mno-avx -msse4
[...@gnu-18 gcc]$ ./a.out 
[...@gnu-18 gcc]$