neon/addhn: initial implementation

meson.build

@@ -14,6 +14,7 @@ simde_neon_families = [
   'abdl',
   'abs',
   'add',
+  'addhn',
   'addl',
   'addlv',
   'addl_high',

simde/arm/neon.h

@@ -34,6 +34,7 @@
 #include "neon/abdl.h"
 #include "neon/abs.h"
 #include "neon/add.h"
+#include "neon/addhn.h"
 #include "neon/addl.h"
 #include "neon/addlv.h"
 #include "neon/addl_high.h"

simde/arm/neon/addhn.h

@@ -0,0 +1,211 @@
+/* SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person
+ * obtaining a copy of this software and associated documentation
+ * files (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy,
+ * modify, merge, publish, distribute, sublicense, and/or sell copies
+ * of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Copyright:
+ *   2021      Evan Nemerson <[email protected]>
+ */
+
+#if !defined(SIMDE_ARM_NEON_ADDHN_H)
+#define SIMDE_ARM_NEON_ADDHN_H
+
+#include "add.h"
+#include "shr_n.h"
+#include "movn.h"
+
+#include "reinterpret.h"
+
+HEDLEY_DIAGNOSTIC_PUSH
+SIMDE_DISABLE_UNWANTED_DIAGNOSTICS
+SIMDE_BEGIN_DECLS_
+
+SIMDE_FUNCTION_ATTRIBUTES
+simde_int8x8_t
+simde_vaddhn_s16(simde_int16x8_t a, simde_int16x8_t b) {
+  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    return vaddhn_s16(a, b);
+  #elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector)
+    simde_int8x8_private r_;
+    simde_int8x16_private tmp_ =
+      simde_int8x16_to_private(
+        simde_vreinterpretq_s8_s16(
+          simde_vaddq_s16(a, b)
+        )
+      );
+    #if SIMDE_ENDIAN_ORDER == SIMDE_ENDIAN_LITTLE
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 1, 3, 5, 7, 9, 11, 13, 15);
+    #else
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 0, 2, 4, 6, 8, 10, 12, 14);
+    #endif
+    return simde_int8x8_from_private(r_);
+  #else
+    return simde_vmovn_s16(simde_vshrq_n_s16(simde_vaddq_s16(a, b), 8));
+  #endif
+}
+#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
+  #undef vaddhn_s16
+  #define vaddhn_s16(a, b) simde_vaddhn_s16((a), (b))
+#endif
+
+SIMDE_FUNCTION_ATTRIBUTES
+simde_int16x4_t
+simde_vaddhn_s32(simde_int32x4_t a, simde_int32x4_t b) {
+  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    return vaddhn_s32(a, b);
+  #elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector)
+    simde_int16x4_private r_;
+    simde_int16x8_private tmp_ =
+      simde_int16x8_to_private(
+        simde_vreinterpretq_s16_s32(
+          simde_vaddq_s32(a, b)
+        )
+      );
+    #if SIMDE_ENDIAN_ORDER == SIMDE_ENDIAN_LITTLE
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 1, 3, 5, 7);
+    #else
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 0, 2, 4, 6);
+    #endif
+    return simde_int16x4_from_private(r_);
+  #else
+    return simde_vmovn_s32(simde_vshrq_n_s32(simde_vaddq_s32(a, b), 16));
+  #endif
+}
+#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
+  #undef vaddhn_s32
+  #define vaddhn_s32(a, b) simde_vaddhn_s32((a), (b))
+#endif
+
+SIMDE_FUNCTION_ATTRIBUTES
+simde_int32x2_t
+simde_vaddhn_s64(simde_int64x2_t a, simde_int64x2_t b) {
+  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    return vaddhn_s64(a, b);
+  #elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector)
+    simde_int32x2_private r_;
+    simde_int32x4_private tmp_ =
+      simde_int32x4_to_private(
+        simde_vreinterpretq_s32_s64(
+          simde_vaddq_s64(a, b)
+        )
+      );
+    #if SIMDE_ENDIAN_ORDER == SIMDE_ENDIAN_LITTLE
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 1, 3);
+    #else
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 0, 2);
+    #endif
+    return simde_int32x2_from_private(r_);
+  #else
+    return simde_vmovn_s64(simde_vshrq_n_s64(simde_vaddq_s64(a, b), 32));
+  #endif
+}
+#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
+  #undef vaddhn_s64
+  #define vaddhn_s64(a, b) simde_vaddhn_s64((a), (b))
+#endif
+
+SIMDE_FUNCTION_ATTRIBUTES
+simde_uint8x8_t
+simde_vaddhn_u16(simde_uint16x8_t a, simde_uint16x8_t b) {
+  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    return vaddhn_u16(a, b);
+  #elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector)
+    simde_uint8x8_private r_;
+    simde_uint8x16_private tmp_ =
+      simde_uint8x16_to_private(
+        simde_vreinterpretq_u8_u16(
+          simde_vaddq_u16(a, b)
+        )
+      );
+    #if SIMDE_ENDIAN_ORDER == SIMDE_ENDIAN_LITTLE
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 1, 3, 5, 7, 9, 11, 13, 15);
+    #else
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 0, 2, 4, 6, 8, 10, 12, 14);
+    #endif
+    return simde_uint8x8_from_private(r_);
+  #else
+    return simde_vmovn_u16(simde_vshrq_n_u16(simde_vaddq_u16(a, b), 8));
+  #endif
+}
+#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
+  #undef vaddhn_u16
+  #define vaddhn_u16(a, b) simde_vaddhn_u16((a), (b))
+#endif
+
+SIMDE_FUNCTION_ATTRIBUTES
+simde_uint16x4_t
+simde_vaddhn_u32(simde_uint32x4_t a, simde_uint32x4_t b) {
+  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    return vaddhn_u32(a, b);
+  #elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector)
+    simde_uint16x4_private r_;
+    simde_uint16x8_private tmp_ =
+      simde_uint16x8_to_private(
+        simde_vreinterpretq_u16_u32(
+          simde_vaddq_u32(a, b)
+        )
+      );
+    #if SIMDE_ENDIAN_ORDER == SIMDE_ENDIAN_LITTLE
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 1, 3, 5, 7);
+    #else
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 0, 2, 4, 6);
+    #endif
+    return simde_uint16x4_from_private(r_);
+  #else
+    return simde_vmovn_u32(simde_vshrq_n_u32(simde_vaddq_u32(a, b), 16));
+  #endif
+}
+#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
+  #undef vaddhn_u32
+  #define vaddhn_u32(a, b) simde_vaddhn_u32((a), (b))
+#endif
+
+SIMDE_FUNCTION_ATTRIBUTES
+simde_uint32x2_t
+simde_vaddhn_u64(simde_uint64x2_t a, simde_uint64x2_t b) {
+  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    return vaddhn_u64(a, b);
+  #elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector)
+    simde_uint32x2_private r_;
+    simde_uint32x4_private tmp_ =
+      simde_uint32x4_to_private(
+        simde_vreinterpretq_u32_u64(
+          simde_vaddq_u64(a, b)
+        )
+      );
+    #if SIMDE_ENDIAN_ORDER == SIMDE_ENDIAN_LITTLE
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 1, 3);
+    #else
+      r_.values = __builtin_shufflevector(tmp_.values, tmp_.values, 0, 2);
+    #endif
+    return simde_uint32x2_from_private(r_);
+  #else
+    return simde_vmovn_u64(simde_vshrq_n_u64(simde_vaddq_u64(a, b), 32));
+  #endif
+}
+#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
+  #undef vaddhn_u64
+  #define vaddhn_u64(a, b) simde_vaddhn_u64((a), (b))
+#endif
+
+SIMDE_END_DECLS_
+HEDLEY_DIAGNOSTIC_POP
+
+#endif /* !defined(SIMDE_ARM_NEON_ADDHN_H) */