https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122778
Bug ID: 122778
Summary: SIMD clone vectorization with AVX512 fails to
implement loop masking
Product: gcc
Version: 16.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: rguenth at gcc dot gnu.org
Target Milestone: ---
diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
index 250c0084d41..f44b4da2321 100644
--- a/gcc/tree-vect-stmts.cc
+++ b/gcc/tree-vect-stmts.cc
@@ -4737,6 +4737,10 @@ vectorizable_simd_clone_call (vec_info *vinfo,
stmt_vec_info stmt_info,
else if (known_eq (atype_subparts,
TYPE_VECTOR_SUBPARTS
(arginfo[i].vectype)))
{
+ /* ??? Not implemented. */
+ gcc_assert (!loop_vinfo
+ || !LOOP_VINFO_FULLY_MASKED_P
(loop_vinfo));
+
/* The vector mask argument matches the input
in the number of lanes, but not necessarily
in the mode. */
No testcase(?).
