[committed,AArch64] Use SVE FABD in conditional arithmetic
diff mbox series

Message ID mptftm2kfqp.fsf@arm.com
State New
Headers show
Series
  • [committed,AArch64] Use SVE FABD in conditional arithmetic
Related show

Commit Message

Richard Sandiford Aug. 15, 2019, 8:14 a.m. UTC
This patch extends the FABD support so that it handles conditional
arithmetic.  We're relying on combine for this, since there's no
associated IFN_COND_* (yet?).

Tested on aarch64-linux-gnu (with and without SVE) and aarch64_be-elf.
Applied as r274507.

Richard


2019-08-15  Richard Sandiford  <richard.sandiford@arm.com>
	    Kugan Vivekanandarajah  <kugan.vivekanandarajah@linaro.org>

gcc/
	* config/aarch64/aarch64-sve.md (*aarch64_cond_abd<SVE_F:mode>_2)
	(*aarch64_cond_abd<SVE_F:mode>_3)
	(*aarch64_cond_abd<SVE_F:mode>_any): New patterns.

gcc/testsuite/
	* gcc.target/aarch64/sve/cond_fabd_1.c: New test.
	* gcc.target/aarch64/sve/cond_fabd_1_run.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_2.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_2_run.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_3.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_3_run.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_4.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_4_run.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_5.c: Likewise.
	* gcc.target/aarch64/sve/cond_fabd_5_run.c: Likewise.

Patch
diff mbox series

Index: gcc/config/aarch64/aarch64-sve.md
===================================================================
--- gcc/config/aarch64/aarch64-sve.md	2019-08-15 09:08:44.073462152 +0100
+++ gcc/config/aarch64/aarch64-sve.md	2019-08-15 09:11:00.736451963 +0100
@@ -2795,6 +2795,123 @@  (define_insn_and_rewrite "*fabd<mode>3"
   }
 )
 
+;; Predicated floating-point absolute difference, merging with the first
+;; input.
+(define_insn_and_rewrite "*aarch64_cond_abd<mode>_2"
+  [(set (match_operand:SVE_F 0 "register_operand" "=w, ?&w")
+	(unspec:SVE_F
+	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl")
+	   (unspec:SVE_F
+	     [(match_operand 4)
+	      (match_operand:SI 5 "aarch64_sve_gp_strictness")
+	      (unspec:SVE_F
+		[(match_operand 6)
+		 (match_operand:SI 7 "aarch64_sve_gp_strictness")
+		 (match_operand:SVE_F 2 "register_operand" "0, w")
+		 (match_operand:SVE_F 3 "register_operand" "w, w")]
+		UNSPEC_COND_FSUB)]
+	     UNSPEC_COND_FABS)
+	   (match_dup 2)]
+	  UNSPEC_SEL))]
+  "TARGET_SVE
+   && aarch64_sve_pred_dominates_p (&operands[4], operands[1])
+   && aarch64_sve_pred_dominates_p (&operands[6], operands[1])"
+  "@
+   fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
+   movprfx\t%0, %2\;fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>"
+  "&& (!rtx_equal_p (operands[1], operands[4])
+       || !rtx_equal_p (operands[1], operands[6]))"
+  {
+    operands[4] = copy_rtx (operands[1]);
+    operands[6] = copy_rtx (operands[1]);
+  }
+  [(set_attr "movprfx" "*,yes")]
+)
+
+;; Predicated floating-point absolute difference, merging with the second
+;; input.
+(define_insn_and_rewrite "*aarch64_cond_abd<mode>_3"
+  [(set (match_operand:SVE_F 0 "register_operand" "=w, ?&w")
+	(unspec:SVE_F
+	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl")
+	   (unspec:SVE_F
+	     [(match_operand 4)
+	      (match_operand:SI 5 "aarch64_sve_gp_strictness")
+	      (unspec:SVE_F
+		[(match_operand 6)
+		 (match_operand:SI 7 "aarch64_sve_gp_strictness")
+		 (match_operand:SVE_F 2 "register_operand" "w, w")
+		 (match_operand:SVE_F 3 "register_operand" "0, w")]
+		UNSPEC_COND_FSUB)]
+	     UNSPEC_COND_FABS)
+	   (match_dup 3)]
+	  UNSPEC_SEL))]
+  "TARGET_SVE
+   && aarch64_sve_pred_dominates_p (&operands[4], operands[1])
+   && aarch64_sve_pred_dominates_p (&operands[6], operands[1])"
+  "@
+   fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %2.<Vetype>
+   movprfx\t%0, %3\;fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %2.<Vetype>"
+  "&& (!rtx_equal_p (operands[1], operands[4])
+       || !rtx_equal_p (operands[1], operands[6]))"
+  {
+    operands[4] = copy_rtx (operands[1]);
+    operands[6] = copy_rtx (operands[1]);
+  }
+  [(set_attr "movprfx" "*,yes")]
+)
+
+;; Predicated floating-point absolute difference, merging with an
+;; independent value.
+(define_insn_and_rewrite "*aarch64_cond_abd<mode>_any"
+  [(set (match_operand:SVE_F 0 "register_operand" "=&w, &w, &w, &w, ?&w")
+	(unspec:SVE_F
+	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl, Upl, Upl, Upl")
+	   (unspec:SVE_F
+	     [(match_operand 5)
+	      (match_operand:SI 6 "aarch64_sve_gp_strictness")
+	      (unspec:SVE_F
+		[(match_operand 7)
+		 (match_operand:SI 8 "aarch64_sve_gp_strictness")
+		 (match_operand:SVE_F 2 "register_operand" "0, w, w, w, w")
+		 (match_operand:SVE_F 3 "register_operand" "w, 0, w, w, w")]
+		UNSPEC_COND_FSUB)]
+	     UNSPEC_COND_FABS)
+	   (match_operand:SVE_F 4 "aarch64_simd_reg_or_zero" "Dz, Dz, Dz, 0, w")]
+	  UNSPEC_SEL))]
+  "TARGET_SVE
+   && !rtx_equal_p (operands[2], operands[4])
+   && !rtx_equal_p (operands[3], operands[4])
+   && aarch64_sve_pred_dominates_p (&operands[5], operands[1])
+   && aarch64_sve_pred_dominates_p (&operands[7], operands[1])"
+  "@
+   movprfx\t%0.<Vetype>, %1/z, %0.<Vetype>\;fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
+   movprfx\t%0.<Vetype>, %1/z, %0.<Vetype>\;fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %2.<Vetype>
+   movprfx\t%0.<Vetype>, %1/z, %2.<Vetype>\;fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
+   movprfx\t%0.<Vetype>, %1/m, %2.<Vetype>\;fabd\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
+   #"
+  "&& 1"
+  {
+    if (reload_completed
+        && register_operand (operands[4], <MODE>mode)
+        && !rtx_equal_p (operands[0], operands[4]))
+      {
+	emit_insn (gen_vcond_mask_<mode><vpred> (operands[0], operands[3],
+						 operands[4], operands[1]));
+	operands[4] = operands[3] = operands[0];
+      }
+    else if (!rtx_equal_p (operands[1], operands[5])
+	     || !rtx_equal_p (operands[1], operands[7]))
+      {
+	operands[5] = copy_rtx (operands[1]);
+	operands[7] = copy_rtx (operands[1]);
+      }
+    else
+      FAIL;
+  }
+  [(set_attr "movprfx" "yes")]
+)
+
 ;; -------------------------------------------------------------------------
 ;; ---- [FP] Multiplication
 ;; -------------------------------------------------------------------------
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_1.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_1.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,29 @@ 
+/* { dg-do compile } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include <stdint.h>
+
+#define DEF_LOOP(TYPE, ABS)				\
+  void __attribute__ ((noinline, noclone))		\
+  test_##TYPE (TYPE *__restrict r, TYPE *__restrict a,	\
+	       TYPE *__restrict b, TYPE *__restrict c,	\
+	       int n)					\
+  {							\
+    for (int i = 0; i < n; ++i)				\
+      r[i] = a[i] < 20 ? ABS (b[i] - c[i]) : b[i];	\
+  }
+
+#define TEST_ALL(T) \
+  T (_Float16, __builtin_fabsf16) \
+  T (float, __builtin_fabsf) \
+  T (double, __builtin_fabs)
+
+TEST_ALL (DEF_LOOP)
+
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.h, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.s, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.d, p[0-7]/m,} 1 } } */
+
+/* { dg-final { scan-assembler-not {\tmov\tz[^,]*z} } } */
+/* { dg-final { scan-assembler-not {\tmovprfx\t} } } */
+/* { dg-final { scan-assembler-not {\tsel\t} } } */
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_1_run.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_1_run.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,33 @@ 
+/* { dg-do run { target aarch64_sve_hw } } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include "cond_fabd_1.c"
+
+#define N 99
+
+#define TEST_LOOP(TYPE, ABS)					\
+  {								\
+    TYPE r[N], a[N], b[N], c[N];				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	a[i] = (i & 1 ? i : 3 * i);				\
+	b[i] = (i >> 4) << (i & 15);				\
+	c[i] = ((i + 2) % 3) * (i + 1);				\
+	asm volatile ("" ::: "memory");				\
+      }								\
+    test_##TYPE (r, a, b, c, N);				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	TYPE expected = a[i] < 20 ? ABS (b[i] - c[i]) : b[i];	\
+	if (r[i] != expected)					\
+	  __builtin_abort ();					\
+	asm volatile ("" ::: "memory");				\
+      }								\
+  }
+
+int
+main (void)
+{
+  TEST_ALL (TEST_LOOP)
+  return 0;
+}
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_2.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_2.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,29 @@ 
+/* { dg-do compile } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include <stdint.h>
+
+#define DEF_LOOP(TYPE, ABS)				\
+  void __attribute__ ((noinline, noclone))		\
+  test_##TYPE (TYPE *__restrict r, TYPE *__restrict a,	\
+	       TYPE *__restrict b, TYPE *__restrict c,	\
+	       int n)					\
+  {							\
+    for (int i = 0; i < n; ++i)				\
+      r[i] = a[i] < 20 ? ABS (b[i] - c[i]) : c[i];	\
+  }
+
+#define TEST_ALL(T) \
+  T (_Float16, __builtin_fabsf16) \
+  T (float, __builtin_fabsf) \
+  T (double, __builtin_fabs)
+
+TEST_ALL (DEF_LOOP)
+
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.h, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.s, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.d, p[0-7]/m,} 1 } } */
+
+/* { dg-final { scan-assembler-not {\tmov\tz[^,]*z} } } */
+/* { dg-final { scan-assembler-not {\tmovprfx\t} } } */
+/* { dg-final { scan-assembler-not {\tsel\t} } } */
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_2_run.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_2_run.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,33 @@ 
+/* { dg-do run { target aarch64_sve_hw } } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include "cond_fabd_2.c"
+
+#define N 99
+
+#define TEST_LOOP(TYPE, ABS)					\
+  {								\
+    TYPE r[N], a[N], b[N], c[N];				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	a[i] = (i & 1 ? i : 3 * i);				\
+	b[i] = (i >> 4) << (i & 15);				\
+	c[i] = ((i + 2) % 3) * (i + 1);				\
+	asm volatile ("" ::: "memory");				\
+      }								\
+    test_##TYPE (r, a, b, c, N);				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	TYPE expected = a[i] < 20 ? ABS (b[i] - c[i]) : c[i];	\
+	if (r[i] != expected)					\
+	  __builtin_abort ();					\
+	asm volatile ("" ::: "memory");				\
+      }								\
+  }
+
+int
+main (void)
+{
+  TEST_ALL (TEST_LOOP)
+  return 0;
+}
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_3.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_3.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,32 @@ 
+/* { dg-do compile } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include <stdint.h>
+
+#define DEF_LOOP(TYPE, ABS)				\
+  void __attribute__ ((noinline, noclone))		\
+  test_##TYPE (TYPE *__restrict r, TYPE *__restrict a,	\
+	       TYPE *__restrict b, TYPE *__restrict c,	\
+	       int n)					\
+  {							\
+    for (int i = 0; i < n; ++i)				\
+      r[i] = a[i] < 20 ? ABS (b[i] - c[i]) : a[i];	\
+  }
+
+#define TEST_ALL(T) \
+  T (_Float16, __builtin_fabsf16) \
+  T (float, __builtin_fabsf) \
+  T (double, __builtin_fabs)
+
+TEST_ALL (DEF_LOOP)
+
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.h, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.s, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.d, p[0-7]/m,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.d, p[0-7]/m, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-not {\tmov\tz[^,]*z} } } */
+/* { dg-final { scan-assembler-not {\tsel\t} } } */
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_3_run.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_3_run.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,33 @@ 
+/* { dg-do run { target aarch64_sve_hw } } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include "cond_fabd_3.c"
+
+#define N 99
+
+#define TEST_LOOP(TYPE, ABS)					\
+  {								\
+    TYPE r[N], a[N], b[N], c[N];				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	a[i] = (i & 1 ? i : 3 * i);				\
+	b[i] = (i >> 4) << (i & 15);				\
+	c[i] = ((i + 2) % 3) * (i + 1);				\
+	asm volatile ("" ::: "memory");				\
+      }								\
+    test_##TYPE (r, a, b, c, N);				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	TYPE expected = a[i] < 20 ? ABS (b[i] - c[i]) : a[i];	\
+	if (r[i] != expected)					\
+	  __builtin_abort ();					\
+	asm volatile ("" ::: "memory");				\
+      }								\
+  }
+
+int
+main (void)
+{
+  TEST_ALL (TEST_LOOP)
+  return 0;
+}
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_4.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_4.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,29 @@ 
+/* { dg-do compile } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include <stdint.h>
+
+#define DEF_LOOP(TYPE, ABS)				\
+  void __attribute__ ((noinline, noclone))		\
+  test_##TYPE (TYPE *__restrict r, TYPE *__restrict a,	\
+	       TYPE *__restrict b, TYPE *__restrict c,	\
+	       int n)					\
+  {							\
+    for (int i = 0; i < n; ++i)				\
+      r[i] = a[i] < 20 ? ABS (b[i] - c[i]) : 8.0;	\
+  }
+
+#define TEST_ALL(T) \
+  T (_Float16, __builtin_fabsf16) \
+  T (float, __builtin_fabsf) \
+  T (double, __builtin_fabs)
+
+TEST_ALL (DEF_LOOP)
+
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.h, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.s, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.d, p[0-7]/m,} 1 } } */
+
+/* { dg-final { scan-assembler-not {\tmov\tz[^,]*z} } } */
+/* { dg-final { scan-assembler-not {\tmovprfx\t} } } */
+/* { dg-final { scan-assembler-times {\tsel\t} 3 } } */
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_4_run.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_4_run.c	2019-08-15 09:11:00.736451963 +0100
@@ -0,0 +1,33 @@ 
+/* { dg-do run { target aarch64_sve_hw } } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include "cond_fabd_4.c"
+
+#define N 99
+
+#define TEST_LOOP(TYPE, ABS)					\
+  {								\
+    TYPE r[N], a[N], b[N], c[N];				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	a[i] = (i & 1 ? i : 3 * i);				\
+	b[i] = (i >> 4) << (i & 15);				\
+	c[i] = ((i + 2) % 3) * (i + 1);				\
+	asm volatile ("" ::: "memory");				\
+      }								\
+    test_##TYPE (r, a, b, c, N);				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	TYPE expected = a[i] < 20 ? ABS (b[i] - c[i]) : 8;	\
+	if (r[i] != expected)					\
+	  __builtin_abort ();					\
+	asm volatile ("" ::: "memory");				\
+      }								\
+  }
+
+int
+main (void)
+{
+  TEST_ALL (TEST_LOOP)
+  return 0;
+}
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_5.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_5.c	2019-08-15 09:11:00.740451935 +0100
@@ -0,0 +1,35 @@ 
+/* { dg-do compile } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include <stdint.h>
+
+#define DEF_LOOP(TYPE, ABS)				\
+  void __attribute__ ((noinline, noclone))		\
+  test_##TYPE (TYPE *__restrict r, TYPE *__restrict a,	\
+	       TYPE *__restrict b, TYPE *__restrict c,	\
+	       int n)					\
+  {							\
+    for (int i = 0; i < n; ++i)				\
+      r[i] = a[i] < 20 ? ABS (b[i] - c[i]) : 0.0;	\
+  }
+
+#define TEST_ALL(T) \
+  T (_Float16, __builtin_fabsf16) \
+  T (float, __builtin_fabsf) \
+  T (double, __builtin_fabs)
+
+TEST_ALL (DEF_LOOP)
+
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.h, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.s, p[0-7]/m,} 1 } } */
+/* { dg-final { scan-assembler-times {\tfabd\tz[0-9]+\.d, p[0-7]/m,} 1 } } */
+
+/* Really we should be able to use MOVPRFX /Z here, but at the moment
+   we're relying on combine to merge a SEL and an arithmetic operation,
+   and the SEL doesn't allow zero operands.  */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.h, p[0-7]/z, z[0-9]+\.h\n} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.s, p[0-7]/z, z[0-9]+\.s\n} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.d, p[0-7]/z, z[0-9]+\.d\n} 1 { xfail *-*-* } } } */
+
+/* { dg-final { scan-assembler-not {\tmov\tz[^,]*z} } } */
+/* { dg-final { scan-assembler-not {\tsel\t} { xfail *-*-* } } } */
Index: gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_5_run.c
===================================================================
--- /dev/null	2019-07-30 08:53:31.317691683 +0100
+++ gcc/testsuite/gcc.target/aarch64/sve/cond_fabd_5_run.c	2019-08-15 09:11:00.740451935 +0100
@@ -0,0 +1,33 @@ 
+/* { dg-do run { target aarch64_sve_hw } } */
+/* { dg-options "-O2 -ftree-vectorize -fno-trapping-math" } */
+
+#include "cond_fabd_5.c"
+
+#define N 99
+
+#define TEST_LOOP(TYPE, ABS)					\
+  {								\
+    TYPE r[N], a[N], b[N], c[N];				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	a[i] = (i & 1 ? i : 3 * i);				\
+	b[i] = (i >> 4) << (i & 15);				\
+	c[i] = ((i + 2) % 3) * (i + 1);				\
+	asm volatile ("" ::: "memory");				\
+      }								\
+    test_##TYPE (r, a, b, c, N);				\
+    for (int i = 0; i < N; ++i)					\
+      {								\
+	TYPE expected = a[i] < 20 ? ABS (b[i] - c[i]) : 0;	\
+	if (r[i] != expected)					\
+	  __builtin_abort ();					\
+	asm volatile ("" ::: "memory");				\
+      }								\
+  }
+
+int
+main (void)
+{
+  TEST_ALL (TEST_LOOP)
+  return 0;
+}