re PR tree-optimization/92645 (Hand written vector code is 450 times slower when compiled with GCC compared to Clang)

2019-11-26  Richard Biener  <rguenther@suse.de>

	PR tree-optimization/92645
	* tree-vect-slp.c (vect_build_slp_tree_2): For unary ops
	do not build the operation from scalars if the operand is.

	* gcc.target/i386/pr92645.c: New testcase.

From-SVN: r278719

4 files changed, 55 insertions, 6 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 186299a..6ea6e5b 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,9 @@
+2019-11-26  Richard Biener  <rguenther@suse.de>
+
+	PR tree-optimization/92645
+	* tree-vect-slp.c (vect_build_slp_tree_2): For unary ops
+	do not build the operation from scalars if the operand is.
+
 2019-11-25  Tobias Burnus  <tobias@codesourcery.com>
 
 	* config/gcn/mkoffload.c (COMMENT_PREFIX, struct id_map,
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 0a284da..d105b60 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2019-11-26  Richard Biener  <rguenther@suse.de>
+
+	PR tree-optimization/92645
+	* gcc.target/i386/pr92645.c: New testcase.
+
 2019-11-26  Jakub Jelinek  <jakub@redhat.com>
 
 	* gfortran.dg/dec-comparison.f90: Change dg-do from run to compile.
diff --git a/gcc/testsuite/gcc.target/i386/pr92645.c b/gcc/testsuite/gcc.target/i386/pr92645.c
new file mode 100644
index 0000000..467ed53
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr92645.c
@@ -0,0 +1,36 @@
+/* { dg-do compile } */
+/* { dg-options "-O3 -fdump-tree-optimized -msse2 -Wno-psabi" } */
+
+typedef unsigned short v8hi __attribute__((vector_size(16)));
+typedef unsigned int v4si __attribute__((vector_size(16)));
+
+void bar (v4si *dst, v8hi * __restrict src)
+{
+  unsigned int tem[8];
+  tem[0] = (*src)[0];
+  tem[1] = (*src)[1];
+  tem[2] = (*src)[2];
+  tem[3] = (*src)[3];
+  tem[4] = (*src)[4];
+  tem[5] = (*src)[5];
+  tem[6] = (*src)[6];
+  tem[7] = (*src)[7];
+  dst[0] = *(v4si *)tem;
+  dst[1] = *(v4si *)&tem[4];
+}
+void foo (v4si *dst, v8hi src)
+{
+  unsigned int tem[8];
+  tem[0] = src[0];
+  tem[1] = src[1];
+  tem[2] = src[2];
+  tem[3] = src[3];
+  tem[4] = src[4];
+  tem[5] = src[5];
+  tem[6] = src[6];
+  tem[7] = src[7];
+  dst[0] = *(v4si *)tem;
+  dst[1] = *(v4si *)&tem[4];
+}
+
+/* { dg-final { scan-tree-dump-times "vec_unpack_" 4 "optimized" } } */
diff --git a/gcc/tree-vect-slp.c b/gcc/tree-vect-slp.c
index bedbe9a..48aca3b 100644
--- a/gcc/tree-vect-slp.c
+++ b/gcc/tree-vect-slp.c
@@ -1410,10 +1410,11 @@ vect_build_slp_tree_2 (vec_info *vinfo,
 					matches, npermutes,
 					&this_tree_size, bst_map)) != NULL)
 	{
-	  /* If we have all children of child built up from scalars then just
-	     throw that away and build it up this node from scalars.  */
+	  /* If we have all children of a non-unary child built up from
+	     scalars then just throw that away and build it up this node
+	     from scalars.  */
 	  if (is_a <bb_vec_info> (vinfo)
-	      && !SLP_TREE_CHILDREN (child).is_empty ()
+	      && SLP_TREE_CHILDREN (child).length () > 1
 	      /* ???  Rejecting patterns this way doesn't work.  We'd have to
 		 do extra work to cancel the pattern so the uses see the
 		 scalar version.  */
@@ -1549,10 +1550,11 @@ vect_build_slp_tree_2 (vec_info *vinfo,
 					    tem, npermutes,
 					    &this_tree_size, bst_map)) != NULL)
 	    {
-	      /* If we have all children of child built up from scalars then
-		 just throw that away and build it up this node from scalars.  */
+	      /* If we have all children of a non-unary child built up from
+		 scalars then just throw that away and build it up this node
+		 from scalars.  */
 	      if (is_a <bb_vec_info> (vinfo)
-		  && !SLP_TREE_CHILDREN (child).is_empty ()
+		  && SLP_TREE_CHILDREN (child).length () > 1
 		  /* ???  Rejecting patterns this way doesn't work.  We'd have
 		     to do extra work to cancel the pattern so the uses see the
 		     scalar version.  */