The following restricts these cases to single-lange SLP as they look
at only the representative scalar argument.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
PR tree-optimization/122826
* tree-vect-stmts.cc (vectorizable_simd_clone_call): Only
use single-lane SLP for SIMD_CLONE_ARG_TYPE_UNIFORM
and SIMD_CLONE_ARG_TYPE_LINEAR_[REF_]CONSTANT_STEP.
---
gcc/tree-vect-stmts.cc | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
index bfed98b8af0..f1870f1c010 100644
--- a/gcc/tree-vect-stmts.cc
+++ b/gcc/tree-vect-stmts.cc
@@ -4159,7 +4159,7 @@ vectorizable_simd_clone_call (vec_info *vinfo,
stmt_vec_info stmt_info,
gcc_assert (thisarginfo.vectype != NULL_TREE);
/* For linear arguments, the analyze phase should have saved
- the base and step in {STMT_VINFO,SLP_TREE}_SIMD_CLONE_INFO. */
+ the base and step. */
if (!cost_vec
&& i * 3 + 4 <= simd_clone_info.length ()
&& simd_clone_info[i * 3 + 2])
@@ -4188,6 +4188,7 @@ vectorizable_simd_clone_call (vec_info *vinfo,
stmt_vec_info stmt_info,
&& thisarginfo.dt != vect_constant_def
&& thisarginfo.dt != vect_external_def
&& loop_vinfo
+ && SLP_TREE_LANES (slp_node) == 1
&& TREE_CODE (op) == SSA_NAME
&& simple_iv (loop, loop_containing_stmt (stmt), op,
&iv, false)
@@ -4198,11 +4199,13 @@ vectorizable_simd_clone_call (vec_info *vinfo,
stmt_vec_info stmt_info,
}
else if ((thisarginfo.dt == vect_constant_def
|| thisarginfo.dt == vect_external_def)
+ && SLP_TREE_LANES (slp_node) == 1
&& POINTER_TYPE_P (TREE_TYPE (op)))
thisarginfo.align = get_pointer_alignment (op) / BITS_PER_UNIT;
/* Addresses of array elements indexed by GOMP_SIMD_LANE are
linear too. */
- if (POINTER_TYPE_P (TREE_TYPE (op))
+ if (SLP_TREE_LANES (slp_node) == 1
+ && POINTER_TYPE_P (TREE_TYPE (op))
&& !thisarginfo.linear_step
&& cost_vec
&& thisarginfo.dt != vect_constant_def
@@ -4279,8 +4282,9 @@ vectorizable_simd_clone_call (vec_info *vinfo,
stmt_vec_info stmt_info,
this_badness += 64;
break;
case SIMD_CLONE_ARG_TYPE_UNIFORM:
- if (arginfo[i].dt != vect_constant_def
- && arginfo[i].dt != vect_external_def)
+ if ((arginfo[i].dt != vect_constant_def
+ && arginfo[i].dt != vect_external_def)
+ || SLP_TREE_LANES (slp_node) != 1)
i = -1;
break;
case SIMD_CLONE_ARG_TYPE_LINEAR_CONSTANT_STEP:
--
2.51.0
