I've had a patch in my dev tree for quite a while that lowers uniform
vector stmts to scalar stmts. This also mitigates PR58497 so I decided
to push it out now.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2015-10-22 Richard Biener <rguent...@suse.de>
PR tree-optimization/58497
* tree-vect-generic.c (ssa_uniform_vector_p): New helper.
(expand_vector_operations_1): Use it. Lower operations on
all uniform vectors to scalar operations if the HW supports it.
* gcc.dg/tree-ssa/vector-5.c: New testcase.
Index: gcc/tree-vect-generic.c
===================================================================
*** gcc/tree-vect-generic.c (revision 229167)
--- gcc/tree-vect-generic.c (working copy)
*************** lower_vec_perm (gimple_stmt_iterator *gs
*** 1339,1344 ****
--- 1339,1361 ----
update_stmt (gsi_stmt (*gsi));
}
+ /* If OP is a uniform vector return the element it is a splat from. */
+
+ static tree
+ ssa_uniform_vector_p (tree op)
+ {
+ if (TREE_CODE (op) == VECTOR_CST
+ || TREE_CODE (op) == CONSTRUCTOR)
+ return uniform_vector_p (op);
+ if (TREE_CODE (op) == SSA_NAME)
+ {
+ gimple *def_stmt = SSA_NAME_DEF_STMT (op);
+ if (gimple_assign_single_p (def_stmt))
+ return uniform_vector_p (gimple_assign_rhs1 (def_stmt));
+ }
+ return NULL_TREE;
+ }
+
/* Return type in which CODE operation with optab OP can be
computed. */
*************** expand_vector_operations_1 (gimple_stmt_
*** 1505,1510 ****
--- 1522,1550 ----
if (TREE_CODE (type) != VECTOR_TYPE)
return;
+ /* If the vector operation is operating on all same vector elements
+ implement it with a scalar operation and a splat if the target
+ supports the scalar operation. */
+ tree srhs1, srhs2 = NULL_TREE;
+ if ((srhs1 = ssa_uniform_vector_p (rhs1)) != NULL_TREE
+ && (rhs2 == NULL_TREE
+ || (srhs2 = ssa_uniform_vector_p (rhs2)) != NULL_TREE)
+ /* As we query direct optabs restrict to non-convert operations. */
+ && TYPE_MODE (TREE_TYPE (type)) == TYPE_MODE (TREE_TYPE (srhs1)))
+ {
+ op = optab_for_tree_code (code, TREE_TYPE (type), optab_scalar);
+ if (optab_handler (op, TYPE_MODE (TREE_TYPE (type))) !=
CODE_FOR_nothing)
+ {
+ tree slhs = make_ssa_name (TREE_TYPE (srhs1));
+ gimple *repl = gimple_build_assign (slhs, code, srhs1, srhs2);
+ gsi_insert_before (gsi, repl, GSI_SAME_STMT);
+ gimple_assign_set_rhs_from_tree (gsi,
+ build_vector_from_val (type, slhs));
+ update_stmt (stmt);
+ return;
+ }
+ }
+
/* A scalar operation pretending to be a vector one. */
if (VECTOR_BOOLEAN_TYPE_P (type)
&& !VECTOR_MODE_P (TYPE_MODE (type))
*************** expand_vector_operations_1 (gimple_stmt_
*** 1554,1568 ****
if (VECTOR_INTEGER_TYPE_P (TREE_TYPE (rhs2)))
{
tree first;
- gimple *def_stmt;
! if ((TREE_CODE (rhs2) == VECTOR_CST
! && (first = uniform_vector_p (rhs2)) != NULL_TREE)
! || (TREE_CODE (rhs2) == SSA_NAME
! && (def_stmt = SSA_NAME_DEF_STMT (rhs2))
! && gimple_assign_single_p (def_stmt)
! && (first = uniform_vector_p
! (gimple_assign_rhs1 (def_stmt))) != NULL_TREE))
{
gimple_assign_set_rhs2 (stmt, first);
update_stmt (stmt);
--- 1594,1601 ----
if (VECTOR_INTEGER_TYPE_P (TREE_TYPE (rhs2)))
{
tree first;
! if ((first = ssa_uniform_vector_p (rhs2)) != NULL_TREE)
{
gimple_assign_set_rhs2 (stmt, first);
update_stmt (stmt);
Index: gcc/testsuite/gcc.dg/tree-ssa/vector-5.c
===================================================================
*** gcc/testsuite/gcc.dg/tree-ssa/vector-5.c (revision 0)
--- gcc/testsuite/gcc.dg/tree-ssa/vector-5.c (working copy)
***************
*** 0 ****
--- 1,15 ----
+ /* { dg-do compile } */
+ /* { dg-options "-O -fdump-tree-optimized" } */
+
+ typedef int v4si __attribute__((vector_size(4*sizeof (int))));
+
+ v4si v;
+ int foo (int i)
+ {
+ v4si v1 = (v4si) { i, i, i, i };
+ v4si v2 = (v4si) { 3, 3, 3, 3 };
+ v = v1 * v2;
+ }
+
+ /* The operation should be carried out as scalar op. */
+ /* { dg-final { scan-tree-dump-times " \* 3;" 1 "optimized" } } */
