aarch64: Add vcopy(q)__lane(q)_bf16 intrinsics

gcc/ChangeLog

2020-10-20  Andrea Corallo  <andrea.corallo@arm.com>

	* config/aarch64/arm_neon.h (vcopy_lane_bf16, vcopyq_lane_bf16)
	(vcopyq_laneq_bf16, vcopy_laneq_bf16): New intrinsics.

gcc/testsuite/ChangeLog

2020-10-20  Andrea Corallo  <andrea.corallo@arm.com>

	* gcc.target/aarch64/advsimd-intrinsics/bf16_vect_copy_lane_1.c:
	New test.
	* gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_2.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_2.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_2.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_2.c:
	Likewise.

10 files changed, 206 insertions, 0 deletions

diff --git a/gcc/config/aarch64/arm_neon.h b/gcc/config/aarch64/arm_neon.h
index 85c0d62..55c2a65 100644
--- a/gcc/config/aarch64/arm_neon.h
+++ b/gcc/config/aarch64/arm_neon.h
@@ -35716,6 +35716,42 @@ vcvtq_high_bf16_f32 (bfloat16x8_t __inactive, float32x4_t __a)
   return __builtin_aarch64_bfcvtn2v8bf (__inactive, __a);
 }
 
+__extension__ extern __inline bfloat16x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+vcopy_lane_bf16 (bfloat16x4_t __a, const int __lane1,
+		 bfloat16x4_t __b, const int __lane2)
+{
+  return __aarch64_vset_lane_any (__aarch64_vget_lane_any (__b, __lane2),
+				  __a, __lane1);
+}
+
+__extension__ extern __inline bfloat16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+vcopyq_lane_bf16 (bfloat16x8_t __a, const int __lane1,
+		  bfloat16x4_t __b, const int __lane2)
+{
+  return __aarch64_vset_lane_any (__aarch64_vget_lane_any (__b, __lane2),
+				  __a, __lane1);
+}
+
+__extension__ extern __inline bfloat16x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+vcopy_laneq_bf16 (bfloat16x4_t __a, const int __lane1,
+		  bfloat16x8_t __b, const int __lane2)
+{
+  return __aarch64_vset_lane_any (__aarch64_vget_lane_any (__b, __lane2),
+				  __a, __lane1);
+}
+
+__extension__ extern __inline bfloat16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+vcopyq_laneq_bf16 (bfloat16x8_t __a, const int __lane1,
+		   bfloat16x8_t __b, const int __lane2)
+{
+  return __aarch64_vset_lane_any (__aarch64_vget_lane_any (__b, __lane2),
+				  __a, __lane1);
+}
+
 #pragma GCC pop_options
 
 /* AdvSIMD 8-bit Integer Matrix Multiply (I8MM) intrinsics.  */
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/bf16_vect_copy_lane_1.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/bf16_vect_copy_lane_1.c
new file mode 100644
index 0000000..d5aa215
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/bf16_vect_copy_lane_1.c
@@ -0,0 +1,32 @@
+/* { dg-do assemble { target { aarch64*-*-* } } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-additional-options "-march=armv8.2-a+bf16 -O3 --save-temps -std=gnu90" } */
+
+#include "arm_neon.h"
+
+bfloat16x4_t __attribute__((noinline,noclone))
+test_vcopy_lane_bf16 (bfloat16x4_t a, bfloat16x4_t b)
+{
+  return vcopy_lane_bf16 (a, 1, b, 2);
+}
+
+bfloat16x8_t __attribute__((noinline,noclone))
+test_vcopyq_lane_bf16 (bfloat16x8_t a, bfloat16x4_t b)
+{
+  return vcopyq_lane_bf16 (a, 1, b, 2);
+}
+
+bfloat16x4_t __attribute__((noinline,noclone))
+test_vcopy_laneq_bf16 (bfloat16x4_t a, bfloat16x8_t b)
+{
+  return vcopy_laneq_bf16 (a, 1, b, 2);
+}
+
+bfloat16x8_t __attribute__((noinline,noclone))
+test_vcopyq_laneq_bf16 (bfloat16x8_t a, bfloat16x8_t b)
+{
+  return vcopyq_laneq_bf16 (a, 1, b, 2);
+}
+
+/* { dg-final { scan-assembler-times "ins\\tv0.h\\\[1\\\], v1.h\\\[2\\\]" 2 } } */
+/* { dg-final { scan-assembler-times "ins\\tv0.h\\\[1\\\], v1.h\\\[0\\\]" 2 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_1.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_1.c
new file mode 100644
index 0000000..7057980
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_1.c
@@ -0,0 +1,18 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x4_t
+test_vcopy_lane_bf16 (bfloat16x4_t a, bfloat16x4_t b)
+{
+  bfloat16x4_t res;
+  res = vcopy_lane_bf16 (a, 0, b, 4);
+  res = vcopy_lane_bf16 (a, 0, b, -1);
+  return res;
+}
+
+/* { dg-error "lane -1 out of range 0 - 3" "" { target *-*-* } 0 } */
+/* { dg-error "lane 4 out of range 0 - 3" "" { target *-*-* } 0 } */
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_2.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_2.c
new file mode 100644
index 0000000..a8ef930
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_lane_bf16_indices_2.c
@@ -0,0 +1,18 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x4_t
+test_vcopy_lane_bf16 (bfloat16x4_t a, bfloat16x4_t b)
+{
+  bfloat16x4_t res;
+  res = vcopy_lane_bf16 (a, -1, b, 2);
+  res = vcopy_lane_bf16 (a, 4, b, 2);
+  return res;
+}
+
+/* { dg-error "lane -1 out of range 0 - 3" "" { target *-*-* } 0 } */
+/* { dg-error "lane 4 out of range 0 - 3" "" { target *-*-* } 0 } */
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_1.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_1.c
new file mode 100644
index 0000000..c156204
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_1.c
@@ -0,0 +1,17 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x4_t
+test_vcopy_laneq_bf16 (bfloat16x4_t a, bfloat16x8_t b)
+{
+  bfloat16x4_t res;
+  /* { dg-error "lane -1 out of range 0 - 3" "" { target *-*-* } 0 } */
+  res = vcopy_laneq_bf16 (a, -1, b, 2);
+  /* { dg-error "lane 4 out of range 0 - 3" "" { target *-*-* } 0 } */
+  res = vcopy_laneq_bf16 (a, 4, b, 2);
+  return res;
+}
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_2.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_2.c
new file mode 100644
index 0000000..036690b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopy_laneq_bf16_indices_2.c
@@ -0,0 +1,17 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x4_t
+test_vcopy_laneq_bf16 (bfloat16x4_t a, bfloat16x8_t b)
+{
+  bfloat16x4_t res;
+  /* { dg-error "lane -1 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopy_laneq_bf16 (a, 1, b, -1);
+  /* { dg-error "lane 8 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopy_laneq_bf16 (a, 1, b, 8);
+  return res;
+}
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_1.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_1.c
new file mode 100644
index 0000000..15fce1b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_1.c
@@ -0,0 +1,17 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x8_t
+test_vcopyq_lane_bf16 (bfloat16x8_t a, bfloat16x4_t b)
+{
+  bfloat16x8_t res;
+  /* { dg-error "lane -1 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopyq_lane_bf16 (a, -1, b, 2);
+  /* { dg-error "lane 8 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopyq_lane_bf16 (a, 8, b, 2);
+  return res;
+}
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_2.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_2.c
new file mode 100644
index 0000000..6e8004a
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_lane_bf16_indices_2.c
@@ -0,0 +1,17 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x8_t
+test_vcopyq_lane_bf16 (bfloat16x8_t a, bfloat16x4_t b)
+{
+  bfloat16x8_t res;
+  /* { dg-error "lane -1 out of range 0 - 3" "" { target *-*-* } 0 } */
+  res = vcopyq_lane_bf16 (a, 2, b, -1);
+  /* { dg-error "lane 4 out of range 0 - 3" "" { target *-*-* } 0 } */
+  res = vcopyq_lane_bf16 (a, 2, b, 4);
+  return res;
+}
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_1.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_1.c
new file mode 100644
index 0000000..2a26b42
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_1.c
@@ -0,0 +1,17 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x8_t
+test_vcopyq_laneq_bf16 (bfloat16x8_t a, bfloat16x8_t b)
+{
+  bfloat16x8_t res;
+  /* { dg-error "lane -1 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopyq_laneq_bf16 (a, -1, b, 2);
+  /* { dg-error "lane 8 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopyq_laneq_bf16 (a, 8, b, 2);
+  return res;
+}
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_2.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_2.c
new file mode 100644
index 0000000..421cb2a
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcopyq_laneq_bf16_indices_2.c
@@ -0,0 +1,17 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon }  */
+
+#include <arm_neon.h>
+
+bfloat16x8_t
+test_vcopyq_laneq_bf16 (bfloat16x8_t a, bfloat16x8_t b)
+{
+  bfloat16x8_t res;
+  /* { dg-error "lane -1 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopyq_laneq_bf16 (a, 2, b, -1);
+  /* { dg-error "lane 8 out of range 0 - 7" "" { target *-*-* } 0 } */
+  res = vcopyq_laneq_bf16 (a, 2, b, 8);
+  return res;
+}