aarch64: Add support for unpacked SVE ABD

This patch adds support for unpacked SVE SABD and UABD.
It also rewrites the patterns so that they match as combine
patterns without the need for REG_EQUAL notes.  Finally,
there was no pattern for merging with the second input,
which can be handled by reversing the operands.

The type suffix needs to be taken from the element size rather
than the container size.

gcc/
	* config/aarch64/aarch64-sve.md (<su>abd<mode>_3): Extend from
	SVE_FULL_I to SVE_I.
	(*aarch64_cond_<su>abd<mode>_2): Likewise.
	(*aarch64_cond_<su>abd<mode>_any): Likewise.
	(@aarch64_pred_<su>abd<mode>): Likewise.  Use UNSPEC_PRED_X
	for the max and min but not for the minus.
	(*aarch64_cond_<su>abd<mode>_3): New pattern.

gcc/testsuite/
	* g++.target/aarch64/sve/abd_1.C: New test.
	* g++.target/aarch64/sve/cond_abd_1.C: Likewise.
	* g++.target/aarch64/sve/cond_abd_2.C: Likewise.
	* g++.target/aarch64/sve/cond_abd_3.C: Likewise.
	* g++.target/aarch64/sve/cond_abd_4.C: Likewise.

6 files changed, 317 insertions, 33 deletions

diff --git a/gcc/config/aarch64/aarch64-sve.md b/gcc/config/aarch64/aarch64-sve.md
index 27b7fd0..8083749 100644
--- a/gcc/config/aarch64/aarch64-sve.md
+++ b/gcc/config/aarch64/aarch64-sve.md
@@ -3973,10 +3973,10 @@
 
 ;; Unpredicated integer absolute difference.
 (define_expand "<su>abd<mode>_3"
-  [(use (match_operand:SVE_FULL_I 0 "register_operand"))
-   (USMAX:SVE_FULL_I
-     (match_operand:SVE_FULL_I 1 "register_operand")
-     (match_operand:SVE_FULL_I 2 "register_operand"))]
+  [(use (match_operand:SVE_I 0 "register_operand"))
+   (USMAX:SVE_I
+     (match_operand:SVE_I 1 "register_operand")
+     (match_operand:SVE_I 2 "register_operand"))]
   "TARGET_SVE"
   {
     rtx pred = aarch64_ptrue_reg (<VPRED>mode);
@@ -3988,17 +3988,20 @@
 
 ;; Predicated integer absolute difference.
 (define_insn "@aarch64_pred_<su>abd<mode>"
-  [(set (match_operand:SVE_FULL_I 0 "register_operand" "=w, ?&w")
-	(unspec:SVE_FULL_I
-	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl")
-	   (minus:SVE_FULL_I
-	     (USMAX:SVE_FULL_I
-	       (match_operand:SVE_FULL_I 2 "register_operand" "%0, w")
-	       (match_operand:SVE_FULL_I 3 "register_operand" "w, w"))
-	     (<max_opp>:SVE_FULL_I
+  [(set (match_operand:SVE_I 0 "register_operand" "=w, ?&w")
+	(minus:SVE_I
+	  (unspec:SVE_I
+	    [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl")
+	     (USMAX:SVE_I
+	       (match_operand:SVE_I 2 "register_operand" "%0, w")
+	       (match_operand:SVE_I 3 "register_operand" "w, w"))]
+	    UNSPEC_PRED_X)
+	  (unspec:SVE_I
+	    [(match_dup 1)
+	     (<max_opp>:SVE_I
 	       (match_dup 2)
-	       (match_dup 3)))]
-	  UNSPEC_PRED_X))]
+	       (match_dup 3))]
+	    UNSPEC_PRED_X)))]
   "TARGET_SVE"
   "@
    <su>abd\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
@@ -4033,19 +4036,19 @@
 
 ;; Predicated integer absolute difference, merging with the first input.
 (define_insn_and_rewrite "*aarch64_cond_<su>abd<mode>_2"
-  [(set (match_operand:SVE_FULL_I 0 "register_operand" "=w, ?&w")
-	(unspec:SVE_FULL_I
+  [(set (match_operand:SVE_I 0 "register_operand" "=w, ?&w")
+	(unspec:SVE_I
 	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl")
-	   (minus:SVE_FULL_I
-	     (unspec:SVE_FULL_I
+	   (minus:SVE_I
+	     (unspec:SVE_I
 	       [(match_operand 4)
-		(USMAX:SVE_FULL_I
-		  (match_operand:SVE_FULL_I 2 "register_operand" "0, w")
-		  (match_operand:SVE_FULL_I 3 "register_operand" "w, w"))]
+		(USMAX:SVE_I
+		  (match_operand:SVE_I 2 "register_operand" "0, w")
+		  (match_operand:SVE_I 3 "register_operand" "w, w"))]
 	       UNSPEC_PRED_X)
-	     (unspec:SVE_FULL_I
+	     (unspec:SVE_I
 	       [(match_operand 5)
-		(<max_opp>:SVE_FULL_I
+		(<max_opp>:SVE_I
 		  (match_dup 2)
 		  (match_dup 3))]
 	       UNSPEC_PRED_X))
@@ -4062,25 +4065,56 @@
   [(set_attr "movprfx" "*,yes")]
 )
 
+;; Predicated integer absolute difference, merging with the second input.
+(define_insn_and_rewrite "*aarch64_cond_<su>abd<mode>_3"
+  [(set (match_operand:SVE_I 0 "register_operand" "=w, ?&w")
+	(unspec:SVE_I
+	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl")
+	   (minus:SVE_I
+	     (unspec:SVE_I
+	       [(match_operand 4)
+		(USMAX:SVE_I
+		  (match_operand:SVE_I 2 "register_operand" "w, w")
+		  (match_operand:SVE_I 3 "register_operand" "0, w"))]
+	       UNSPEC_PRED_X)
+	     (unspec:SVE_I
+	       [(match_operand 5)
+		(<max_opp>:SVE_I
+		  (match_dup 2)
+		  (match_dup 3))]
+	       UNSPEC_PRED_X))
+	   (match_dup 3)]
+	  UNSPEC_SEL))]
+  "TARGET_SVE"
+  "@
+   <su>abd\t%0.<Vetype>, %1/m, %0.<Vetype>, %2.<Vetype>
+   movprfx\t%0, %3\;<su>abd\t%0.<Vetype>, %1/m, %0.<Vetype>, %2.<Vetype>"
+  "&& (!CONSTANT_P (operands[4]) || !CONSTANT_P (operands[5]))"
+  {
+    operands[4] = operands[5] = CONSTM1_RTX (<VPRED>mode);
+  }
+  [(set_attr "movprfx" "*,yes")]
+)
+
 ;; Predicated integer absolute difference, merging with an independent value.
 (define_insn_and_rewrite "*aarch64_cond_<su>abd<mode>_any"
-  [(set (match_operand:SVE_FULL_I 0 "register_operand" "=&w, &w, &w, &w, ?&w")
-	(unspec:SVE_FULL_I
+  [(set (match_operand:SVE_I 0 "register_operand" "=&w, &w, &w, &w, ?&w")
+	(unspec:SVE_I
 	  [(match_operand:<VPRED> 1 "register_operand" "Upl, Upl, Upl, Upl, Upl")
-	   (minus:SVE_FULL_I
-	     (unspec:SVE_FULL_I
+	   (minus:SVE_I
+	     (unspec:SVE_I
 	       [(match_operand 5)
-		(USMAX:SVE_FULL_I
-		  (match_operand:SVE_FULL_I 2 "register_operand" "0, w, w, w, w")
-		  (match_operand:SVE_FULL_I 3 "register_operand" "w, 0, w, w, w"))]
+		(USMAX:SVE_I
+		  (match_operand:SVE_I 2 "register_operand" "0, w, w, w, w")
+		  (match_operand:SVE_I 3 "register_operand" "w, 0, w, w, w"))]
 	       UNSPEC_PRED_X)
-	     (unspec:SVE_FULL_I
+	     (unspec:SVE_I
 	       [(match_operand 6)
-		(<max_opp>:SVE_FULL_I
+		(<max_opp>:SVE_I
 		  (match_dup 2)
 		  (match_dup 3))]
 	       UNSPEC_PRED_X))
-	   (match_operand:SVE_FULL_I 4 "aarch64_simd_reg_or_zero" "Dz, Dz, Dz, 0, w")]
+	   (match_operand:SVE_I 4 "aarch64_simd_reg_or_zero" "Dz, Dz, Dz, 0, w")]
 	  UNSPEC_SEL))]
   "TARGET_SVE
    && !rtx_equal_p (operands[2], operands[4])
diff --git a/gcc/testsuite/g++.target/aarch64/sve/abd_1.C b/gcc/testsuite/g++.target/aarch64/sve/abd_1.C
new file mode 100644
index 0000000..0facf58
--- /dev/null
+++ b/gcc/testsuite/g++.target/aarch64/sve/abd_1.C
@@ -0,0 +1,38 @@
+/* { dg-do assemble { target aarch64_asm_sve_ok } } */
+/* { dg-options "-O -msve-vector-bits=2048 -save-temps" } */
+
+#include <stdint.h>
+
+#define TEST_OP(TYPE) \
+  TYPE \
+  test##_##TYPE##_reg (TYPE a, TYPE b) \
+  { \
+    return (a > b ? a : b) - (a < b ? a : b); \
+  }
+
+#define TEST_TYPE(TYPE, SIZE) \
+  typedef TYPE TYPE##SIZE __attribute__((vector_size(SIZE))); \
+  TEST_OP (TYPE##SIZE)
+
+TEST_TYPE (int8_t, 32)
+TEST_TYPE (uint8_t, 32)
+
+TEST_TYPE (int8_t, 64)
+TEST_TYPE (uint8_t, 64)
+TEST_TYPE (int16_t, 64)
+TEST_TYPE (uint16_t, 64)
+
+TEST_TYPE (int8_t, 128)
+TEST_TYPE (uint8_t, 128)
+TEST_TYPE (int16_t, 128)
+TEST_TYPE (uint16_t, 128)
+TEST_TYPE (int32_t, 128)
+TEST_TYPE (uint32_t, 128)
+
+/* { dg-final { scan-assembler-times {\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, z[0-9]+\.b\n} 3 } } */
+/* { dg-final { scan-assembler-times {\tsabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 2 } } */
+/* { dg-final { scan-assembler-times {\tsabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, z[0-9]+\.b\n} 3 } } */
+/* { dg-final { scan-assembler-times {\tuabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 2 } } */
+/* { dg-final { scan-assembler-times {\tuabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
diff --git a/gcc/testsuite/g++.target/aarch64/sve/cond_abd_1.C b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_1.C
new file mode 100644
index 0000000..a93d1d1
--- /dev/null
+++ b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_1.C
@@ -0,0 +1,60 @@
+/* { dg-do assemble { target aarch64_asm_sve_ok } } */
+/* { dg-options "-O -msve-vector-bits=2048 -save-temps" } */
+
+#include <stdint.h>
+
+#define TEST_OP(TYPE) \
+  TYPE \
+  test##_##TYPE##_reg (TYPE a, TYPE b, TYPE c) \
+  { \
+    return c == 0 ? (a > b ? a : b) - (a < b ? a : b) : a; \
+  }
+
+#define TEST_TYPE(TYPE, SIZE) \
+  typedef TYPE TYPE##SIZE __attribute__((vector_size(SIZE))); \
+  TEST_OP (TYPE##SIZE)
+
+TEST_TYPE (int8_t, 32)
+TEST_TYPE (uint8_t, 32)
+
+TEST_TYPE (int8_t, 64)
+TEST_TYPE (uint8_t, 64)
+TEST_TYPE (int16_t, 64)
+TEST_TYPE (uint16_t, 64)
+
+TEST_TYPE (int8_t, 128)
+TEST_TYPE (uint8_t, 128)
+TEST_TYPE (int16_t, 128)
+TEST_TYPE (uint16_t, 128)
+TEST_TYPE (int32_t, 128)
+TEST_TYPE (uint32_t, 128)
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x0\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tsabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tsabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tsabd\t\1\.s, p[0-7]/m, \1\.s, z[0-9]+\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x1\]\n[^L]*\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tsabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tsabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tsabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, \1\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x0\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tuabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tuabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tuabd\t\1\.s, p[0-7]/m, \1\.s, z[0-9]+\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x1\]\n[^L]*\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tuabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tuabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tuabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, \1\.s\n} } } */
+
+/* { dg-final { scan-assembler-not {\tsel\t} } } */
diff --git a/gcc/testsuite/g++.target/aarch64/sve/cond_abd_2.C b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_2.C
new file mode 100644
index 0000000..ad6304a
--- /dev/null
+++ b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_2.C
@@ -0,0 +1,60 @@
+/* { dg-do assemble { target aarch64_asm_sve_ok } } */
+/* { dg-options "-O -msve-vector-bits=2048 -save-temps" } */
+
+#include <stdint.h>
+
+#define TEST_OP(TYPE) \
+  TYPE \
+  test##_##TYPE##_reg (TYPE a, TYPE b, TYPE c) \
+  { \
+    return c == 0 ? (a > b ? a : b) - (a < b ? a : b) : b; \
+  }
+
+#define TEST_TYPE(TYPE, SIZE) \
+  typedef TYPE TYPE##SIZE __attribute__((vector_size(SIZE))); \
+  TEST_OP (TYPE##SIZE)
+
+TEST_TYPE (int8_t, 32)
+TEST_TYPE (uint8_t, 32)
+
+TEST_TYPE (int8_t, 64)
+TEST_TYPE (uint8_t, 64)
+TEST_TYPE (int16_t, 64)
+TEST_TYPE (uint16_t, 64)
+
+TEST_TYPE (int8_t, 128)
+TEST_TYPE (uint8_t, 128)
+TEST_TYPE (int16_t, 128)
+TEST_TYPE (uint16_t, 128)
+TEST_TYPE (int32_t, 128)
+TEST_TYPE (uint32_t, 128)
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x1\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tsabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tsabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tsabd\t\1\.s, p[0-7]/m, \1\.s, z[0-9]+\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x0\]\n[^L]*\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tsabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tsabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tsabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, \1\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x1\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x1\]\n[^L]*\tuabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tuabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x1\]\n[^L]*\tuabd\t\1\.s, p[0-7]/m, \1\.s, z[0-9]+\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x0\]\n[^L]*\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, \1\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x0\]\n[^L]*\tuabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tuabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, \1\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x0\]\n[^L]*\tuabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, \1\.s\n} } } */
+
+/* { dg-final { scan-assembler-not {\tsel\t} } } */
diff --git a/gcc/testsuite/g++.target/aarch64/sve/cond_abd_3.C b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_3.C
new file mode 100644
index 0000000..29a8d2f
--- /dev/null
+++ b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_3.C
@@ -0,0 +1,49 @@
+/* { dg-do assemble { target aarch64_asm_sve_ok } } */
+/* { dg-options "-O -msve-vector-bits=2048 -save-temps" } */
+
+#include <stdint.h>
+
+#define TEST_OP(TYPE) \
+  TYPE \
+  test##_##TYPE##_reg (TYPE a, TYPE b, TYPE c) \
+  { \
+    return c == 0 ? (a > b ? a : b) - (a < b ? a : b) : c; \
+  }
+
+#define TEST_TYPE(TYPE, SIZE) \
+  typedef TYPE TYPE##SIZE __attribute__((vector_size(SIZE))); \
+  TEST_OP (TYPE##SIZE)
+
+TEST_TYPE (int8_t, 32)
+TEST_TYPE (uint8_t, 32)
+
+TEST_TYPE (int8_t, 64)
+TEST_TYPE (uint8_t, 64)
+TEST_TYPE (int16_t, 64)
+TEST_TYPE (uint16_t, 64)
+
+TEST_TYPE (int8_t, 128)
+TEST_TYPE (uint8_t, 128)
+TEST_TYPE (int16_t, 128)
+TEST_TYPE (uint16_t, 128)
+TEST_TYPE (int32_t, 128)
+TEST_TYPE (uint32_t, 128)
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x2\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x2\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x2\]\n[^L]*\tsabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x2\]\n[^L]*\tsabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x2\]\n[^L]*\tsabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x2\]\n[^L]*\tsabd\t\1\.s, p[0-7]/m, \1\.s, z[0-9]+\.s\n} } } */
+
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.h, p[0-7]/z, \[x2\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.s, p[0-7]/z, \[x2\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1b\t(z[0-9]+)\.d, p[0-7]/z, \[x2\]\n[^L]*\tuabd\t\1\.b, p[0-7]/m, \1\.b, z[0-9]+\.b\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.s, p[0-7]/z, \[x2\]\n[^L]*\tuabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1h\t(z[0-9]+)\.d, p[0-7]/z, \[x2\]\n[^L]*\tuabd\t\1\.h, p[0-7]/m, \1\.h, z[0-9]+\.h\n} } } */
+/* { dg-final { scan-assembler {\tld1w\t(z[0-9]+)\.d, p[0-7]/z, \[x2\]\n[^L]*\tuabd\t\1\.s, p[0-7]/m, \1\.s, z[0-9]+\.s\n} } } */
+
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b\n} 6 } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h\n} 4 } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s\n} 2 } } */
+/* { dg-final { scan-assembler-not {\tsel\t} } } */
diff --git a/gcc/testsuite/g++.target/aarch64/sve/cond_abd_4.C b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_4.C
new file mode 100644
index 0000000..b415731
--- /dev/null
+++ b/gcc/testsuite/g++.target/aarch64/sve/cond_abd_4.C
@@ -0,0 +1,43 @@
+/* { dg-do assemble { target aarch64_asm_sve_ok } } */
+/* { dg-options "-O -msve-vector-bits=2048 -save-temps" } */
+
+#include <stdint.h>
+
+#define TEST_OP(TYPE) \
+  TYPE \
+  test##_##TYPE##_reg (TYPE a, TYPE b, TYPE c) \
+  { \
+    return c == 1 ? (a > b ? a : b) - (a < b ? a : b) : 0; \
+  }
+
+#define TEST_TYPE(TYPE, SIZE) \
+  typedef TYPE TYPE##SIZE __attribute__((vector_size(SIZE))); \
+  TEST_OP (TYPE##SIZE)
+
+TEST_TYPE (int8_t, 32)
+TEST_TYPE (uint8_t, 32)
+
+TEST_TYPE (int8_t, 64)
+TEST_TYPE (uint8_t, 64)
+TEST_TYPE (int16_t, 64)
+TEST_TYPE (uint16_t, 64)
+
+TEST_TYPE (int8_t, 128)
+TEST_TYPE (uint8_t, 128)
+TEST_TYPE (int16_t, 128)
+TEST_TYPE (uint16_t, 128)
+TEST_TYPE (int32_t, 128)
+TEST_TYPE (uint32_t, 128)
+
+/* { dg-final { scan-assembler-times {\tsabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, z[0-9]+\.b\n} 3 } } */
+/* { dg-final { scan-assembler-times {\tsabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 2 } } */
+/* { dg-final { scan-assembler-times {\tsabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tuabd\tz[0-9]+\.b, p[0-7]/m, z[0-9]+\.b, z[0-9]+\.b\n} 3 } } */
+/* { dg-final { scan-assembler-times {\tuabd\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 2 } } */
+/* { dg-final { scan-assembler-times {\tuabd\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.b, p[0-7]/z, z[0-9]+\.b\n} 6 } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.h, p[0-7]/z, z[0-9]+\.h\n} 4 } } */
+/* { dg-final { scan-assembler-times {\tmovprfx\tz[0-9]+\.s, p[0-7]/z, z[0-9]+\.s\n} 2 } } */
+/* { dg-final { scan-assembler-not {\tsel\t} } } */