[PATCH] match.pd: fold the inequality spelling of divisibility

From: Kyrylo Tkachov <[email protected]>

GCC folds:

  x / y * y == x

to:

  x % y == 0

Apply the same simplification to inequality.

  int divisible_by_3 (unsigned int x)
  {
    return x / 3 * 3 != x;
  }

aarch64 -O2:

before:

	divisible_by_3:
		mov	w1, 43691
		movk	w1, 0xaaaa, lsl 16
		umull	x1, w0, w1
		lsr	x1, x1, 33
		add	w1, w1, w1, lsl 1
		cmp	w1, w0
		cset	w0, ne
		ret

after:

	divisible_by_3:
		mov	w1, 43691
		movk	w1, 0xaaaa, lsl 16
		mul	w0, w0, w1
		mov	w1, 1431655765
		cmp	w0, w1
		cset	w0, hi
		ret

Bootstrapped and tested on aarch64-none-linux-gnu.
Ok for trunk?
Thanks,
Kyrill

gcc/ChangeLog:

	* match.pd (x / y * y == x): Extend to inequality.

gcc/testsuite/ChangeLog:

	* gcc.dg/tree-ssa/divmul-mod-1.c: New test.

Signed-off-by: Kyrylo Tkachov <[email protected]>
---
 gcc/match.pd                                 |  9 ++---
 gcc/testsuite/gcc.dg/tree-ssa/divmul-mod-1.c | 38 ++++++++++++++++++++
 2 files changed, 43 insertions(+), 4 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/divmul-mod-1.c

diff --git a/gcc/match.pd b/gcc/match.pd
index 877e449d7d8..33eec29f53a 100644
--- a/gcc/match.pd
+++ b/gcc/match.pd
@@ -6036,16 +6036,17 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
    (plus:c (mult:c (div @0 @1) @1) (mod @0 @1))
    @0))
 
-/* x / y * y == x -> x % y == 0.  */
-(simplify
-  (eq:c (mult:c (trunc_div:s @0 @1) @1) @0)
+/* x / y * y == x -> x % y == 0, and likewise for !=.  */
+(for cmp (eq ne)
+ (simplify
+  (cmp:c (mult:c (trunc_div:s @0 @1) @1) @0)
   (if (TREE_CODE (TREE_TYPE (@0)) != COMPLEX_TYPE
        && (!VECTOR_MODE_P (TYPE_MODE (TREE_TYPE (@0)))
 	   || !target_supports_op_p (TREE_TYPE (@0), TRUNC_DIV_EXPR,
 				     optab_vector)
 	   || target_supports_op_p (TREE_TYPE (@0), TRUNC_MOD_EXPR,
 				    optab_vector)))
-   (eq (trunc_mod @0 @1) { build_zero_cst (TREE_TYPE (@0)); })))
+   (cmp (trunc_mod @0 @1) { build_zero_cst (TREE_TYPE (@0)); }))))
 
 /* ((X /[ex] C1) +- C2) * (C1 * C3)  -->  (X * C3) +- (C1 * C2 * C3).  */
 (for op (plus minus)
diff --git a/gcc/testsuite/gcc.dg/tree-ssa/divmul-mod-1.c b/gcc/testsuite/gcc.dg/tree-ssa/divmul-mod-1.c
new file mode 100644
index 00000000000..f65a28d2335
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/tree-ssa/divmul-mod-1.c
@@ -0,0 +1,38 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -fdump-tree-optimized" } */
+
+typedef int v4si __attribute__ ((vector_size (16)));
+typedef unsigned int v4ui __attribute__ ((vector_size (16)));
+
+int
+eq (unsigned int a, unsigned int b)
+{
+  return a / b * b == a;
+}
+
+int
+ne (unsigned int a, unsigned int b)
+{
+  return a / b * b != a;
+}
+
+int
+nes (int a, int b)
+{
+  return a / b * b != a;
+}
+
+v4si
+veq (v4ui a, v4ui b)
+{
+  return a / b * b == a;
+}
+
+v4si
+vne (v4ui a, v4ui b)
+{
+  return a / b * b != a;
+}
+
+/* { dg-final { scan-tree-dump-not " \\* " "optimized" } } */
+/* { dg-final { scan-tree-dump " % " "optimized" } } */
-- 
2.50.1 (Apple Git-155)