tree-optimization/124555 - guard abs patterns for vector support

The following adds missing vector support checks to abs producing
match.pd patterns.  The g++.dg/absvect.C shows that we previously
then lowered this to non-vector unless you add -msse4 on x86_64
at least.

	PR tree-optimization/124555
	* match.pd ((A - B) >=/> 0 ? (A - B) : (B - A) -> abs (A - B)):
	Guard the vector case with target_supports_op_p checks.
	((A - B) <=/< 0 ? (A - B) : (B - A) -> -abs (A - B)): Likewise.
	((type)A >=/> 0 ? A : -A -> abs (A)): Likewise.
	((type)A <=/< 0 ? A : -A -> -abs (A)): Likewise.

	* gcc.dg/torture/pr124555.c: New testcase.
	* g++.dg/absvect.C: Restrict dump scan to x86-64 and force
	-msse4 there.

diff --git a/gcc/match.pd b/gcc/match.pd
index 1d6428bf7e53e8849e22cb3595d324f431a3c0ab..7b652afb43de4bc10c729a48cc32ae6766dd247e 100644 (file)
--- a/gcc/match.pd
+++ b/gcc/match.pd
@@ -7237,7 +7237,9 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
         && !TYPE_UNSIGNED (TREE_TYPE (@1))
         && element_precision (@1)
              <= element_precision (@0)
-        && bitwise_equal_p (@1, @2))
+        && bitwise_equal_p (@1, @2)
+        && (!VECTOR_TYPE_P (type)
+            || target_supports_op_p (type, ABS_EXPR, optab_vector)))
     (if (TYPE_UNSIGNED (TREE_TYPE (@2)))
       (with {
        tree stype = signed_type_for (TREE_TYPE (@2));
@@ -7255,7 +7257,10 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
         && !TYPE_UNSIGNED (TREE_TYPE (@1))
         && element_precision (@1)
              <= element_precision (@0)
-        && bitwise_equal_p (@1, @2))
+        && bitwise_equal_p (@1, @2)
+        && (!VECTOR_TYPE_P (type)
+            || (target_supports_op_p (type, ABS_EXPR, optab_vector)
+                && target_supports_op_p (type, NEGATE_EXPR, optab_vector))))
       (if ((ANY_INTEGRAL_TYPE_P (TREE_TYPE (@2))
              && !TYPE_OVERFLOW_WRAPS (TREE_TYPE (@2)))
            || TYPE_UNSIGNED (TREE_TYPE (@2)))
@@ -7294,7 +7299,9 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
   (simplify
    (cnd (cmp (minus@0 @1 @2) zerop) @0 (minus @2 @1))
    (if (!HONOR_SIGNED_ZEROS (type)
-       && !TYPE_UNSIGNED (type))
+       && !TYPE_UNSIGNED (type)
+       && (!VECTOR_TYPE_P (type)
+           || target_supports_op_p (type, ABS_EXPR, optab_vector)))
     (abs @0))))
  /* (A - B) <=/< 0 ? (A - B) : (B - A)    same as -abs (A - B) */
  (for cmp (le lt)
@@ -7303,7 +7310,10 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
    (if (!HONOR_SIGNED_ZEROS (type)
        && !TYPE_UNSIGNED (type))
     (if (ANY_INTEGRAL_TYPE_P (type)
-        && !TYPE_OVERFLOW_WRAPS (type))
+        && !TYPE_OVERFLOW_WRAPS (type)
+        && (!VECTOR_TYPE_P (type)
+            || (target_supports_op_p (type, ABS_EXPR, optab_vector)
+                && target_supports_op_p (type, NEGATE_EXPR, optab_vector))))
      (with {
         tree utype = unsigned_type_for (type);
       }

diff --git a/gcc/testsuite/g++.dg/absvect.C b/gcc/testsuite/g++.dg/absvect.C
index 5cf2ca307f43b67d60b32959ec931b2a02f7d590..d28ea846d74d2fb941c59ff146b29b61e6fae508 100644 (file)
--- a/gcc/testsuite/g++.dg/absvect.C
+++ b/gcc/testsuite/g++.dg/absvect.C
@@ -1,6 +1,7 @@
 /* { dg-do compile } */
 /* { dg-options "-Ofast -fdump-tree-phiopt1" } */
-/* { dg-final { scan-tree-dump-times " = ABS_EXPR <x_\[0-9]*\\\(D\\\)>;" 1 "phiopt1" } } */
+/* { dg-additional-options "-msse4" { target { x86_64-*-* i?86-*-* } } } */
+/* { dg-final { scan-tree-dump-times " = ABS_EXPR <x_\[0-9]*\\\(D\\\)>;" 1 "phiopt1" { target { x86_64-*-* i?86-*-* } } } } */
 
 typedef int v2si __attribute__ ((vector_size (2 * sizeof(int))));
 typedef short v2hi __attribute__ ((vector_size (2 * sizeof(short))));

diff --git a/gcc/testsuite/gcc.dg/torture/pr124555.c b/gcc/testsuite/gcc.dg/torture/pr124555.c
new file mode 100644 (file)
index 0000000..668ad7b
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/torture/pr124555.c
@@ -0,0 +1,25 @@
+/* { dg-additional-options "-fwrapv" } */
+
+int vsad8_c_y;
+char *vsad8_c_s1, *vsad8_c_s2;
+long vsad8_c_stride;
+int vsad8_c()
+{
+  int score, x;
+  for (; vsad8_c_y; vsad8_c_y++)
+    {
+      x = 0;
+      for (; x < 8; x++)
+       score += (vsad8_c_s1[x] - vsad8_c_s2[x] - vsad8_c_s1[x + vsad8_c_stride]
+           + vsad8_c_s2[x + vsad8_c_stride]) >= 0
+           ? vsad8_c_s1[x] - vsad8_c_s2[x] -
+           vsad8_c_s1[x + vsad8_c_stride] +
+           vsad8_c_s2[x + vsad8_c_stride]
+           : -(vsad8_c_s1[x] - vsad8_c_s2[x] -
+               vsad8_c_s1[x + vsad8_c_stride] +
+               vsad8_c_s2[x + vsad8_c_stride]);
+      vsad8_c_s1 += vsad8_c_stride;
+      vsad8_c_s2 += vsad8_c_stride;
+    }
+  return score;
+}