Instead of going via the PHI node def, use the scalar reduction
input from the reduction stmt.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
Sorry for the re-post, this is now re-based on pristine trunk,
hopefully making git am happy.
* tree-vect-loop.cc (vectorize_fold_left_reduction): Get
reduc_var as argument.
(vect_transform_reduction): Adjust.
---
gcc/tree-vect-loop.cc | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
index f22613e5a3a..f48a18a593d 100644
--- a/gcc/tree-vect-loop.cc
+++ b/gcc/tree-vect-loop.cc
@@ -6501,7 +6501,7 @@ vectorize_fold_left_reduction (loop_vec_info loop_vinfo,
stmt_vec_info stmt_info,
gimple_stmt_iterator *gsi,
slp_tree slp_node,
- gimple *reduc_def_stmt,
+ tree reduc_var,
code_helper code, internal_fn reduc_fn,
int num_ops, tree vectype_in,
int reduc_index, vec_loop_masks *masks,
@@ -6546,7 +6546,6 @@ vectorize_fold_left_reduction (loop_vec_info loop_vinfo,
gimple *sdef = vect_orig_stmt (scalar_dest_def_info)->stmt;
tree scalar_dest = gimple_get_lhs (sdef);
tree scalar_type = TREE_TYPE (scalar_dest);
- tree reduc_var = gimple_phi_result (reduc_def_stmt);
int vec_num = vec_oprnds0.length ();
tree vec_elem_type = TREE_TYPE (vectype_out);
@@ -8016,8 +8015,6 @@ vect_transform_reduction (loop_vec_info loop_vinfo,
The last use is the reduction variable. In case of nested cycle this
assumption is not true: we use reduc_index to record the index of the
reduction variable. */
- stmt_vec_info phi_info = STMT_VINFO_REDUC_DEF (vect_orig_stmt (stmt_info));
- gphi *reduc_def_phi = as_a <gphi *> (phi_info->stmt);
int reduc_index = STMT_VINFO_REDUC_IDX (stmt_info);
tree vectype_in = SLP_TREE_VECTYPE (SLP_TREE_CHILDREN (slp_node)[0]);
@@ -8059,8 +8056,9 @@ vect_transform_reduction (loop_vec_info loop_vinfo,
{
internal_fn reduc_fn = STMT_VINFO_REDUC_FN (reduc_info);
gcc_assert (code.is_tree_code () || cond_fn_p);
+ tree reduc_var = op.ops[reduc_index];
return vectorize_fold_left_reduction
- (loop_vinfo, stmt_info, gsi, slp_node, reduc_def_phi,
+ (loop_vinfo, stmt_info, gsi, slp_node, reduc_var,
code, reduc_fn, op.num_ops, vectype_in,
reduc_index, masks, lens);
}
--
2.43.0
