https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123063
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |missed-optimization
Component|middle-end |rtl-optimization
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
gimple level looks good:
```
ivtmp.34_12 = (unsigned long) &v_copy;
_9 = ivtmp.34_12 + 400;
<bb 3> [local count: 265751102]:
# vect_res_12.27_16 = PHI <vect_res_8.28_20(3), { 0, 0, 0, 0 }(2)>
# ivtmp.34_3 = PHI <ivtmp.34_14(3), ivtmp.34_12(2)>
_1 = (void *) ivtmp.34_3;
vect__1.26_15 = MEM <vector(4) int> [(int *)_1];
vect_res_8.29_17 = VIEW_CONVERT_EXPR<vector(4) unsigned int>(vect__1.26_15);
vect_res_8.29_18 = VIEW_CONVERT_EXPR<vector(4) unsigned
int>(vect_res_12.27_16);
vect_res_8.29_19 = vect_res_8.29_17 + vect_res_8.29_18;
vect_res_8.28_20 = VIEW_CONVERT_EXPR<vector(4) int>(vect_res_8.29_19);
ivtmp.34_14 = ivtmp.34_3 + 16;
if (_9 != ivtmp.34_14)
goto <bb 3>; [95.96%]
else
goto <bb 4>; [4.04%]
```