aarch64: Add codegen support for SVE2 faminmax

The AArch64 FEAT_FAMINMAX extension introduces instructions for
computing the floating point absolute maximum and minimum of the
two vectors element-wise.

This patch adds code generation for famax and famin in terms of existing
unspecs. With this patch:
1. famax can be expressed as taking UNSPEC_COND_SMAX of the two operands
   and then taking absolute value of their result.
2. famin can be expressed as taking UNSPEC_COND_SMIN of the two operands
   and then taking absolute value of their result.

This fusion of operators is only possible when
-march=armv9-a+faminmax+sve flags are passed. We also need to pass
-ffast-math flag; this is what enables compiler to use UNSPEC_COND_SMAX
and UNSPEC_COND_SMIN.

This code generation is only available on -O2 or -O3 as that is when
auto-vectorization is enabled.

gcc/ChangeLog:

	* config/aarch64/aarch64-sve2.md
	(*aarch64_pred_faminmax_fused): Instruction pattern for faminmax
	codegen.
	* config/aarch64/iterators.md: Iterator and attribute for
	faminmax codegen.

gcc/testsuite/ChangeLog:

	* gcc.target/aarch64/sve/faminmax_1.c: New test.
	* gcc.target/aarch64/sve/faminmax_2.c: New test.

4 files changed, 147 insertions, 0 deletions

diff --git a/gcc/config/aarch64/aarch64-sve2.md b/gcc/config/aarch64/aarch64-sve2.md
index 725092c..5f2697c 100644
--- a/gcc/config/aarch64/aarch64-sve2.md
+++ b/gcc/config/aarch64/aarch64-sve2.md
@@ -2467,6 +2467,43 @@
   [(set_attr "movprfx" "yes")]
 )
 
+;; -------------------------------------------------------------------------
+;; -- [FP] Absolute maximum and minimum
+;; -------------------------------------------------------------------------
+;; Includes:
+;; - FAMAX
+;; - FAMIN
+;; -------------------------------------------------------------------------
+;; Predicated floating-point absolute maximum and minimum.
+(define_insn_and_rewrite "*aarch64_pred_faminmax_fused"
+  [(set (match_operand:SVE_FULL_F 0 "register_operand")
+	(unspec:SVE_FULL_F
+	  [(match_operand:<VPRED> 1 "register_operand")
+	   (match_operand:SI 4 "aarch64_sve_gp_strictness")
+	   (unspec:SVE_FULL_F
+	     [(match_operand 5)
+	      (const_int SVE_RELAXED_GP)
+	      (match_operand:SVE_FULL_F 2 "register_operand")]
+	     UNSPEC_COND_FABS)
+	   (unspec:SVE_FULL_F
+	     [(match_operand 6)
+	      (const_int SVE_RELAXED_GP)
+	      (match_operand:SVE_FULL_F 3 "register_operand")]
+	     UNSPEC_COND_FABS)]
+	  SVE_COND_SMAXMIN))]
+  "TARGET_SVE_FAMINMAX"
+  {@ [ cons: =0 , 1   , 2  , 3 ; attrs: movprfx ]
+     [ w        , Upl , %0 , w ; *              ] <faminmax_cond_uns_op>\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
+     [ ?&w      , Upl , w  , w ; yes            ] movprfx\t%0, %2\;<faminmax_cond_uns_op>\t%0.<Vetype>, %1/m, %0.<Vetype>, %3.<Vetype>
+  }
+  "&& (!rtx_equal_p (operands[1], operands[5])
+       || !rtx_equal_p (operands[1], operands[6]))"
+  {
+    operands[5] = copy_rtx (operands[1]);
+    operands[6] = copy_rtx (operands[1]);
+  }
+)
+
 ;; =========================================================================
 ;; == Complex arithmetic
 ;; =========================================================================
diff --git a/gcc/config/aarch64/iterators.md b/gcc/config/aarch64/iterators.md
index a04f9f9..efba783 100644
--- a/gcc/config/aarch64/iterators.md
+++ b/gcc/config/aarch64/iterators.md
@@ -3142,6 +3142,9 @@
 					 UNSPEC_COND_SMAX
 					 UNSPEC_COND_SMIN])
 
+(define_int_iterator SVE_COND_SMAXMIN [UNSPEC_COND_SMAX
+				       UNSPEC_COND_SMIN])
+
 (define_int_iterator SVE_COND_FP_TERNARY [UNSPEC_COND_FMLA
 					  UNSPEC_COND_FMLS
 					  UNSPEC_COND_FNMLA
@@ -4502,6 +4505,9 @@
 
 (define_int_iterator FAMINMAX_UNS [UNSPEC_FAMAX UNSPEC_FAMIN])
 
+(define_int_attr faminmax_cond_uns_op
+  [(UNSPEC_COND_SMAX "famax") (UNSPEC_COND_SMIN "famin")])
+
 (define_int_attr faminmax_uns_op
   [(UNSPEC_FAMAX "famax") (UNSPEC_FAMIN "famin")])
 
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/faminmax_1.c b/gcc/testsuite/gcc.target/aarch64/sve/faminmax_1.c
new file mode 100644
index 0000000..3b65cce
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/faminmax_1.c
@@ -0,0 +1,44 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-O3 -ffast-math" } */
+
+#include "arm_sve.h"
+
+#pragma GCC target "+sve+faminmax"
+
+#define TEST_FAMAX(TYPE)						\
+  void fn_famax_##TYPE (TYPE * restrict a,				\
+			TYPE * restrict b,				\
+			TYPE * restrict c,				\
+			int n) {					\
+    for (int i = 0; i < n; i++) {					\
+      TYPE temp1 = __builtin_fabs (a[i]);				\
+      TYPE temp2 = __builtin_fabs (b[i]);				\
+      c[i] = __builtin_fmax (temp1, temp2);				\
+    }									\
+  }									\
+
+#define TEST_FAMIN(TYPE)						\
+  void fn_famin_##TYPE (TYPE * restrict a,				\
+			TYPE * restrict b,				\
+			TYPE * restrict c,				\
+			int n) {					\
+    for (int i = 0; i < n; i++) {					\
+      TYPE temp1 = __builtin_fabs (a[i]);				\
+      TYPE temp2 = __builtin_fabs (b[i]);				\
+      c[i] = __builtin_fmin (temp1, temp2);				\
+    }									\
+  }									\
+
+TEST_FAMAX (float16_t)
+TEST_FAMAX (float32_t)
+TEST_FAMAX (float64_t)
+TEST_FAMIN (float16_t)
+TEST_FAMIN (float32_t)
+TEST_FAMIN (float64_t)
+
+/* { dg-final { scan-assembler-times {\tfamax\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfamax\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfamax\tz[0-9]+\.d, p[0-7]/m, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfamin\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfamin\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfamin\tz[0-9]+\.d, p[0-7]/m, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/faminmax_2.c b/gcc/testsuite/gcc.target/aarch64/sve/faminmax_2.c
new file mode 100644
index 0000000..d80f6ec
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/faminmax_2.c
@@ -0,0 +1,60 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-O3 -ffast-math" } */
+
+#include "arm_sve.h"
+
+#pragma GCC target "+sve+faminmax"
+
+#define TEST_WITH_SVMAX(TYPE)						\
+  TYPE fn_fmax_##TYPE (TYPE x, TYPE y) {				\
+    svbool_t pg = svptrue_b8();						\
+    return svmax_x(pg, svabs_x(pg, x), svabs_x(pg, y));			\
+  }									\
+
+#define TEST_WITH_SVMAXNM(TYPE)						\
+  TYPE fn_fmaxnm_##TYPE (TYPE x, TYPE y) {				\
+    svbool_t pg = svptrue_b8();						\
+    return svmaxnm_x(pg, svabs_x(pg, x), svabs_x(pg, y));		\
+  }									\
+
+#define TEST_WITH_SVMIN(TYPE)						\
+  TYPE fn_fmin_##TYPE (TYPE x, TYPE y) {				\
+    svbool_t pg = svptrue_b8();						\
+    return svmin_x(pg, svabs_x(pg, x), svabs_x(pg, y));			\
+  }									\
+
+#define TEST_WITH_SVMINNM(TYPE)						\
+  TYPE fn_fminnm_##TYPE (TYPE x, TYPE y) {				\
+    svbool_t pg = svptrue_b8();						\
+    return svminnm_x(pg, svabs_x(pg, x), svabs_x(pg, y));		\
+  }									\
+
+TEST_WITH_SVMAX (svfloat16_t)
+TEST_WITH_SVMAX (svfloat32_t)
+TEST_WITH_SVMAX (svfloat64_t)
+
+TEST_WITH_SVMAXNM (svfloat16_t)
+TEST_WITH_SVMAXNM (svfloat32_t)
+TEST_WITH_SVMAXNM (svfloat64_t)
+
+TEST_WITH_SVMIN (svfloat16_t)
+TEST_WITH_SVMIN (svfloat32_t)
+TEST_WITH_SVMIN (svfloat64_t)
+
+TEST_WITH_SVMINNM (svfloat16_t)
+TEST_WITH_SVMINNM (svfloat32_t)
+TEST_WITH_SVMINNM (svfloat64_t)
+
+/* { dg-final { scan-assembler-not {\tfamax\t} } }  */
+/* { dg-final { scan-assembler-not {\tfamin\t} } }  */
+
+/* { dg-final { scan-assembler-times {\tfabs\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h\n} 8 } } */
+/* { dg-final { scan-assembler-times {\tfabs\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s\n} 8 } } */
+/* { dg-final { scan-assembler-times {\tfabs\tz[0-9]+\.d, p[0-7]/m, z[0-9]+\.d\n} 8 } } */
+
+/* { dg-final { scan-assembler-times {\tfmax\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfmax\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfmax\tz[0-9]+\.d, p[0-7]/m, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfmin\tz[0-9]+\.h, p[0-7]/m, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfmin\tz[0-9]+\.s, p[0-7]/m, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tfmin\tz[0-9]+\.d, p[0-7]/m, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */