[Bug middle-end/126410] vectorizer aligns 16-byte array to 64byte

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126410

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot 
gnu.org

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
The issue is that we call ensure_base_align when costing only.  Fix:

diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
index 488b9754bed..c5a86d3a1a4 100644
--- a/gcc/tree-vect-stmts.cc
+++ b/gcc/tree-vect-stmts.cc
@@ -8304,7 +8304,8 @@ vectorizable_store (vec_info *vinfo,

   /* Transform.  */

-  ensure_base_align (dr_info);
+  if (!costing_p)
+    ensure_base_align (dr_info);

   if (STMT_VINFO_SIMD_LANE_ACCESS_P (stmt_info) >= 3)
     {
@@ -9986,7 +9987,8 @@ vectorizable_load (vec_info *vinfo,
   /* Transform.  */

   dr_vec_info *dr_info = STMT_VINFO_DR_INFO (stmt_info), *first_dr_info =
NULL;
-  ensure_base_align (dr_info);
+  if (!costing_p)
+    ensure_base_align (dr_info);

   if (memory_access_type == VMAT_INVARIANT)
     {