re PR tree-optimization/81303 (410.bwaves regression caused by r249919)

2017-12-08  Richard Biener  <rguenther@suse.de>

	PR tree-optimization/81303
	* tree-vect-stmts.c (vect_is_simple_cond): For invariant
	conditions try to create a comparison vector type matching
	the data vector type.
	(vectorizable_condition): Adjust.
	* tree-vect-patterns.c (vect_recog_mask_conversion_pattern):
	Leave invariant conditions alone in case we can vectorize those.

	* gcc.target/i386/vectorize9.c: New testcase.
	* gcc.target/i386/vectorize10.c: New testcase.

From-SVN: r255497

6 files changed, 91 insertions, 2 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 1d20dc5..a666d11 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,13 @@
+2017-12-08  Richard Biener  <rguenther@suse.de>
+
+	PR tree-optimization/81303
+	* tree-vect-stmts.c (vect_is_simple_cond): For invariant
+	conditions try to create a comparison vector type matching
+	the data vector type.
+	(vectorizable_condition): Adjust.
+	* tree-vect-patterns.c (vect_recog_mask_conversion_pattern):
+	Leave invariant conditions alone in case we can vectorize those.
+
 2017-12-08  Julia Koval  <julia.koval@intel.com>
 
 	* config/i386/avx512vnniintrin.h (_mm512_dpwssd_epi32,
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index b13130d..8324d4f 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,9 @@
+2017-12-08  Richard Biener  <rguenther@suse.de>
+
+	PR tree-optimization/81303
+	* gcc.target/i386/vectorize9.c: New testcase.
+	* gcc.target/i386/vectorize10.c: New testcase.
+
 2017-12-08  Julia Koval  <julia.koval@intel.com>
 
 	* gcc.target/i386/avx512f-vnni-1.c: Add vdpwssd checks.
diff --git a/gcc/testsuite/gcc.target/i386/vectorize10.c b/gcc/testsuite/gcc.target/i386/vectorize10.c
new file mode 100644
index 0000000..4353aec
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/vectorize10.c
@@ -0,0 +1,16 @@
+/* { dg-do compile } */
+/* { dg-options "-O -ftree-loop-vectorize -fdump-tree-vect-details -fno-tree-loop-im -msse2 -mno-avx" } */
+
+double x[1024][1024], red[1024];
+void foo (void)
+{
+  for (int i = 0; i < 1024; ++i)
+    for (int j = 0; j < 1024; ++j)
+      {
+	double v = i == 0 ? 0.0 : red[j];
+	v = v + x[i][j];
+	red[j] = v;
+      }
+}
+
+/* { dg-final { scan-tree-dump "vectorized 1 loops" "vect" } } */
diff --git a/gcc/testsuite/gcc.target/i386/vectorize9.c b/gcc/testsuite/gcc.target/i386/vectorize9.c
new file mode 100644
index 0000000..3f02be4
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/vectorize9.c
@@ -0,0 +1,16 @@
+/* { dg-do compile } */
+/* { dg-options "-O -ftree-loop-vectorize -fdump-tree-vect-details -fno-tree-loop-im -mavx2 -mprefer-vector-width=256" } */
+
+double x[1024][1024], red[1024];
+void foo (void)
+{
+  for (int i = 0; i < 1024; ++i)
+    for (int j = 0; j < 1024; ++j)
+      {
+	double v = i == 0 ? 0.0 : red[j];
+	v = v + x[i][j];
+	red[j] = v;
+      }
+}
+
+/* { dg-final { scan-tree-dump "vectorization factor = 4" "vect" } } */
diff --git a/gcc/tree-vect-patterns.c b/gcc/tree-vect-patterns.c
index e4051b6..a2c6293 100644
--- a/gcc/tree-vect-patterns.c
+++ b/gcc/tree-vect-patterns.c
@@ -3976,6 +3976,32 @@ vect_recog_mask_conversion_pattern (vec<gimple *> *stmts, tree *type_in,
 	  || TYPE_VECTOR_SUBPARTS (vectype1) == TYPE_VECTOR_SUBPARTS (vectype2))
 	return NULL;
 
+      /* If rhs1 is invariant and we can promote it leave the COND_EXPR
+         in place, we can handle it in vectorizable_condition.  This avoids
+	 unnecessary promotion stmts and increased vectorization factor.  */
+      if (COMPARISON_CLASS_P (rhs1)
+	  && INTEGRAL_TYPE_P (rhs1_type)
+	  && TYPE_VECTOR_SUBPARTS (vectype1) < TYPE_VECTOR_SUBPARTS (vectype2))
+	{
+	  gimple *dummy;
+	  enum vect_def_type dt;
+	  if (vect_is_simple_use (TREE_OPERAND (rhs1, 0), stmt_vinfo->vinfo,
+				  &dummy, &dt)
+	      && dt == vect_external_def
+	      && vect_is_simple_use (TREE_OPERAND (rhs1, 1), stmt_vinfo->vinfo,
+				     &dummy, &dt)
+	      && (dt == vect_external_def
+		  || dt == vect_constant_def))
+	    {
+	      tree wide_scalar_type = build_nonstandard_integer_type
+		(tree_to_uhwi (TYPE_SIZE (TREE_TYPE (vectype1))),
+		 TYPE_UNSIGNED (rhs1_type));
+	      tree vectype3 = get_vectype_for_scalar_type (wide_scalar_type);
+	      if (expand_vec_cond_expr_p (vectype1, vectype3, TREE_CODE (rhs1)))
+		return NULL;
+	    }
+	}
+
       /* If rhs1 is a comparison we need to move it into a
 	 separate statement.  */
       if (TREE_CODE (rhs1) != SSA_NAME)
diff --git a/gcc/tree-vect-stmts.c b/gcc/tree-vect-stmts.c
index ee2f1ca..2bebad1 100644
--- a/gcc/tree-vect-stmts.c
+++ b/gcc/tree-vect-stmts.c
@@ -7791,7 +7791,8 @@ vectorizable_load (gimple *stmt, gimple_stmt_iterator *gsi, gimple **vec_stmt,
 
 static bool
 vect_is_simple_cond (tree cond, vec_info *vinfo,
-		     tree *comp_vectype, enum vect_def_type *dts)
+		     tree *comp_vectype, enum vect_def_type *dts,
+		     tree vectype)
 {
   tree lhs, rhs;
   tree vectype1 = NULL_TREE, vectype2 = NULL_TREE;
@@ -7844,6 +7845,20 @@ vect_is_simple_cond (tree cond, vec_info *vinfo,
     return false;
 
   *comp_vectype = vectype1 ? vectype1 : vectype2;
+  /* Invariant comparison.  */
+  if (! *comp_vectype)
+    {
+      tree scalar_type = TREE_TYPE (lhs);
+      /* If we can widen the comparison to match vectype do so.  */
+      if (INTEGRAL_TYPE_P (scalar_type)
+	  && tree_int_cst_lt (TYPE_SIZE (scalar_type),
+			      TYPE_SIZE (TREE_TYPE (vectype))))
+	scalar_type = build_nonstandard_integer_type
+	  (tree_to_uhwi (TYPE_SIZE (TREE_TYPE (vectype))),
+	   TYPE_UNSIGNED (scalar_type));
+      *comp_vectype = get_vectype_for_scalar_type (scalar_type);
+    }
+
   return true;
 }
 
@@ -7941,7 +7956,7 @@ vectorizable_condition (gimple *stmt, gimple_stmt_iterator *gsi,
   else_clause = gimple_assign_rhs3 (stmt);
 
   if (!vect_is_simple_cond (cond_expr, stmt_info->vinfo,
-			    &comp_vectype, &dts[0])
+			    &comp_vectype, &dts[0], vectype)
       || !comp_vectype)
     return false;