https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91861
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Known to fail| |10.0
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
C testcase
float x[4], y[4];
int r[4];
void foo()
{
for (int i = 0; i < 4; ++i)
r[i] = __builtin_isless (x[i], y[i]);
}
vectorizes to
vect__1.5_17 = MEM <vector(4) float> [(float *)&x];
vect__2.8_12 = MEM <vector(4) float> [(float *)&y];
vect_patt_21.11_24 = VEC_COND_EXPR <vect__2.8_12 u<= vect__1.5_17, { 0, 0, 0,
0 }, { 1, 1, 1, 1 }>;
MEM <vector(4) int> [(int *)&r] = vect_patt_21.11_24;
which I think is OK.
