https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81635
--- Comment #3 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #2)
> So maybe finally a testcase where that SCEV analysis did sth useful...
>
> x86_64 testcase should be possible with changing the datatype to double?
Yep.
This fails to vectorize with -O2 -ftree-slp-vectorize:
....
double p[1000] __attribute__((aligned(8)));
double p2[1000] __attribute__((aligned(8)));
void __attribute__((noinline, noclone))
foo ()
{
double a, b;
unsigned int i;
for (i = 0; i < 1000; i += 2)
{
a = p[i];
b = p[i+1];
p2[i] = a;
p2[i+1] = b;
}
}
...
Changing 'i' to type int allows it to be vectorized.
