https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61680
--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
/* When we perform grouped accesses and perform implicit CSE
by detecting equal accesses and doing disambiguation with
runtime alias tests like for
.. = a[i];
.. = a[i+1];
a[i] = ..;
a[i+1] = ..;
*p = ..;
.. = a[i];
.. = a[i+1];
where we will end up loading { a[i], a[i+1] } once, make
sure that inserting group loads before the first load and
stores after the last store will do the right thing. */
if ((STMT_VINFO_GROUPED_ACCESS (stmtinfo_a)
&& GROUP_SAME_DR_STMT (stmtinfo_a))
|| (STMT_VINFO_GROUPED_ACCESS (stmtinfo_b)
&& GROUP_SAME_DR_STMT (stmtinfo_b)))
{
gimple earlier_stmt;
earlier_stmt = get_earlier_stmt (DR_STMT (dra), DR_STMT (drb));
if (DR_IS_WRITE
(STMT_VINFO_DATA_REF (vinfo_for_stmt (earlier_stmt))))
{
if (dump_enabled_p ())
dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
"READ_WRITE dependence in interleaving."
"\n");
return true;
}
}
is supposed to catch this kind of issue ... but it's very simple-minded
and GROUP_SAME_DR_STMT is not set (we don't have redundant loads to CSE).
What is interesting is that we don't CSE the loads from w[i_26][{0,1}].
That would have likely "fixed" this as well ...
Removing the GROUP_SAME_DR_STMT tests fixes the testcase.