------- Comment #35 from rguenth at gcc dot gnu dot org 2008-10-04 15:58
-------
Another missed optimization on the tree level is hoisting of the load of
array[k*4] before the k < j condition which is possible after the PRE
insertion:
<bb 3>:
if (k_4 < j_5(D))
goto <bb 4>;
else
goto <bb 12>;
<bb 12>:
pretmp.12_31 = k_4 * 4;
pretmp.13_30 = array_8(D) + pretmp.12_31;
pretmp.14_25 = *pretmp.13_30;
goto <bb 6>;
<bb 4>:
D.1242_7 = k_4 * 4;
D.1243_9 = array_8(D) + D.1242_7;
D.1244_10 = *D.1243_9;
maybe PRE could somehow even insert on BB3 entry instead of BB3 exit edges in
this case. This would re-enable if-conversion of the second branch.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21485
