xref: /prebuilts/sdk/renderscript/clang-include/arm_neon.h

/*===---- arm_neon.h - ARM Neon intrinsics ---------------------------------===
 *
 * 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.
 *
 *===-----------------------------------------------------------------------===
 */

#ifndef __ARM_NEON_H
#define __ARM_NEON_H

#if !defined(__ARM_NEON)
#error "NEON support not enabled"
#endif

#include <stdint.h>

typedef float float32_t;
typedef __fp16 float16_t;
#ifdef __aarch64__
typedef double float64_t;
#endif

#ifdef __aarch64__
typedef uint8_t poly8_t;
typedef uint16_t poly16_t;
typedef uint64_t poly64_t;
typedef __uint128_t poly128_t;
#else
typedef int8_t poly8_t;
typedef int16_t poly16_t;
#endif
typedef __attribute__((neon_vector_type(8))) int8_t int8x8_t;
typedef __attribute__((neon_vector_type(16))) int8_t int8x16_t;
typedef __attribute__((neon_vector_type(4))) int16_t int16x4_t;
typedef __attribute__((neon_vector_type(8))) int16_t int16x8_t;
typedef __attribute__((neon_vector_type(2))) int32_t int32x2_t;
typedef __attribute__((neon_vector_type(4))) int32_t int32x4_t;
typedef __attribute__((neon_vector_type(1))) int64_t int64x1_t;
typedef __attribute__((neon_vector_type(2))) int64_t int64x2_t;
typedef __attribute__((neon_vector_type(8))) uint8_t uint8x8_t;
typedef __attribute__((neon_vector_type(16))) uint8_t uint8x16_t;
typedef __attribute__((neon_vector_type(4))) uint16_t uint16x4_t;
typedef __attribute__((neon_vector_type(8))) uint16_t uint16x8_t;
typedef __attribute__((neon_vector_type(2))) uint32_t uint32x2_t;
typedef __attribute__((neon_vector_type(4))) uint32_t uint32x4_t;
typedef __attribute__((neon_vector_type(1))) uint64_t uint64x1_t;
typedef __attribute__((neon_vector_type(2))) uint64_t uint64x2_t;
typedef __attribute__((neon_vector_type(4))) float16_t float16x4_t;
typedef __attribute__((neon_vector_type(8))) float16_t float16x8_t;
typedef __attribute__((neon_vector_type(2))) float32_t float32x2_t;
typedef __attribute__((neon_vector_type(4))) float32_t float32x4_t;
#ifdef __aarch64__
typedef __attribute__((neon_vector_type(1))) float64_t float64x1_t;
typedef __attribute__((neon_vector_type(2))) float64_t float64x2_t;
#endif
typedef __attribute__((neon_polyvector_type(8))) poly8_t poly8x8_t;
typedef __attribute__((neon_polyvector_type(16))) poly8_t poly8x16_t;
typedef __attribute__((neon_polyvector_type(4))) poly16_t poly16x4_t;
typedef __attribute__((neon_polyvector_type(8))) poly16_t poly16x8_t;
#ifdef __aarch64__
typedef __attribute__((neon_polyvector_type(1))) poly64_t poly64x1_t;
typedef __attribute__((neon_polyvector_type(2))) poly64_t poly64x2_t;
#endif

typedef struct int8x8x2_t {
  int8x8_t val[2];
} int8x8x2_t;

typedef struct int8x16x2_t {
  int8x16_t val[2];
} int8x16x2_t;

typedef struct int16x4x2_t {
  int16x4_t val[2];
} int16x4x2_t;

typedef struct int16x8x2_t {
  int16x8_t val[2];
} int16x8x2_t;

typedef struct int32x2x2_t {
  int32x2_t val[2];
} int32x2x2_t;

typedef struct int32x4x2_t {
  int32x4_t val[2];
} int32x4x2_t;

typedef struct int64x1x2_t {
  int64x1_t val[2];
} int64x1x2_t;

typedef struct int64x2x2_t {
  int64x2_t val[2];
} int64x2x2_t;

typedef struct uint8x8x2_t {
  uint8x8_t val[2];
} uint8x8x2_t;

typedef struct uint8x16x2_t {
  uint8x16_t val[2];
} uint8x16x2_t;

typedef struct uint16x4x2_t {
  uint16x4_t val[2];
} uint16x4x2_t;

typedef struct uint16x8x2_t {
  uint16x8_t val[2];
} uint16x8x2_t;

typedef struct uint32x2x2_t {
  uint32x2_t val[2];
} uint32x2x2_t;

typedef struct uint32x4x2_t {
  uint32x4_t val[2];
} uint32x4x2_t;

typedef struct uint64x1x2_t {
  uint64x1_t val[2];
} uint64x1x2_t;

typedef struct uint64x2x2_t {
  uint64x2_t val[2];
} uint64x2x2_t;

typedef struct float16x4x2_t {
  float16x4_t val[2];
} float16x4x2_t;

typedef struct float16x8x2_t {
  float16x8_t val[2];
} float16x8x2_t;

typedef struct float32x2x2_t {
  float32x2_t val[2];
} float32x2x2_t;

typedef struct float32x4x2_t {
  float32x4_t val[2];
} float32x4x2_t;

#ifdef __aarch64__
typedef struct float64x1x2_t {
  float64x1_t val[2];
} float64x1x2_t;

typedef struct float64x2x2_t {
  float64x2_t val[2];
} float64x2x2_t;

#endif
typedef struct poly8x8x2_t {
  poly8x8_t val[2];
} poly8x8x2_t;

typedef struct poly8x16x2_t {
  poly8x16_t val[2];
} poly8x16x2_t;

typedef struct poly16x4x2_t {
  poly16x4_t val[2];
} poly16x4x2_t;

typedef struct poly16x8x2_t {
  poly16x8_t val[2];
} poly16x8x2_t;

#ifdef __aarch64__
typedef struct poly64x1x2_t {
  poly64x1_t val[2];
} poly64x1x2_t;

typedef struct poly64x2x2_t {
  poly64x2_t val[2];
} poly64x2x2_t;

#endif
typedef struct int8x8x3_t {
  int8x8_t val[3];
} int8x8x3_t;

typedef struct int8x16x3_t {
  int8x16_t val[3];
} int8x16x3_t;

typedef struct int16x4x3_t {
  int16x4_t val[3];
} int16x4x3_t;

typedef struct int16x8x3_t {
  int16x8_t val[3];
} int16x8x3_t;

typedef struct int32x2x3_t {
  int32x2_t val[3];
} int32x2x3_t;

typedef struct int32x4x3_t {
  int32x4_t val[3];
} int32x4x3_t;

typedef struct int64x1x3_t {
  int64x1_t val[3];
} int64x1x3_t;

typedef struct int64x2x3_t {
  int64x2_t val[3];
} int64x2x3_t;

typedef struct uint8x8x3_t {
  uint8x8_t val[3];
} uint8x8x3_t;

typedef struct uint8x16x3_t {
  uint8x16_t val[3];
} uint8x16x3_t;

typedef struct uint16x4x3_t {
  uint16x4_t val[3];
} uint16x4x3_t;

typedef struct uint16x8x3_t {
  uint16x8_t val[3];
} uint16x8x3_t;

typedef struct uint32x2x3_t {
  uint32x2_t val[3];
} uint32x2x3_t;

typedef struct uint32x4x3_t {
  uint32x4_t val[3];
} uint32x4x3_t;

typedef struct uint64x1x3_t {
  uint64x1_t val[3];
} uint64x1x3_t;

typedef struct uint64x2x3_t {
  uint64x2_t val[3];
} uint64x2x3_t;

typedef struct float16x4x3_t {
  float16x4_t val[3];
} float16x4x3_t;

typedef struct float16x8x3_t {
  float16x8_t val[3];
} float16x8x3_t;

typedef struct float32x2x3_t {
  float32x2_t val[3];
} float32x2x3_t;

typedef struct float32x4x3_t {
  float32x4_t val[3];
} float32x4x3_t;

#ifdef __aarch64__
typedef struct float64x1x3_t {
  float64x1_t val[3];
} float64x1x3_t;

typedef struct float64x2x3_t {
  float64x2_t val[3];
} float64x2x3_t;

#endif
typedef struct poly8x8x3_t {
  poly8x8_t val[3];
} poly8x8x3_t;

typedef struct poly8x16x3_t {
  poly8x16_t val[3];
} poly8x16x3_t;

typedef struct poly16x4x3_t {
  poly16x4_t val[3];
} poly16x4x3_t;

typedef struct poly16x8x3_t {
  poly16x8_t val[3];
} poly16x8x3_t;

#ifdef __aarch64__
typedef struct poly64x1x3_t {
  poly64x1_t val[3];
} poly64x1x3_t;

typedef struct poly64x2x3_t {
  poly64x2_t val[3];
} poly64x2x3_t;

#endif
typedef struct int8x8x4_t {
  int8x8_t val[4];
} int8x8x4_t;

typedef struct int8x16x4_t {
  int8x16_t val[4];
} int8x16x4_t;

typedef struct int16x4x4_t {
  int16x4_t val[4];
} int16x4x4_t;

typedef struct int16x8x4_t {
  int16x8_t val[4];
} int16x8x4_t;

typedef struct int32x2x4_t {
  int32x2_t val[4];
} int32x2x4_t;

typedef struct int32x4x4_t {
  int32x4_t val[4];
} int32x4x4_t;

typedef struct int64x1x4_t {
  int64x1_t val[4];
} int64x1x4_t;

typedef struct int64x2x4_t {
  int64x2_t val[4];
} int64x2x4_t;

typedef struct uint8x8x4_t {
  uint8x8_t val[4];
} uint8x8x4_t;

typedef struct uint8x16x4_t {
  uint8x16_t val[4];
} uint8x16x4_t;

typedef struct uint16x4x4_t {
  uint16x4_t val[4];
} uint16x4x4_t;

typedef struct uint16x8x4_t {
  uint16x8_t val[4];
} uint16x8x4_t;

typedef struct uint32x2x4_t {
  uint32x2_t val[4];
} uint32x2x4_t;

typedef struct uint32x4x4_t {
  uint32x4_t val[4];
} uint32x4x4_t;

typedef struct uint64x1x4_t {
  uint64x1_t val[4];
} uint64x1x4_t;

typedef struct uint64x2x4_t {
  uint64x2_t val[4];
} uint64x2x4_t;

typedef struct float16x4x4_t {
  float16x4_t val[4];
} float16x4x4_t;

typedef struct float16x8x4_t {
  float16x8_t val[4];
} float16x8x4_t;

typedef struct float32x2x4_t {
  float32x2_t val[4];
} float32x2x4_t;

typedef struct float32x4x4_t {
  float32x4_t val[4];
} float32x4x4_t;

#ifdef __aarch64__
typedef struct float64x1x4_t {
  float64x1_t val[4];
} float64x1x4_t;

typedef struct float64x2x4_t {
  float64x2_t val[4];
} float64x2x4_t;

#endif
typedef struct poly8x8x4_t {
  poly8x8_t val[4];
} poly8x8x4_t;

typedef struct poly8x16x4_t {
  poly8x16_t val[4];
} poly8x16x4_t;

typedef struct poly16x4x4_t {
  poly16x4_t val[4];
} poly16x4x4_t;

typedef struct poly16x8x4_t {
  poly16x8_t val[4];
} poly16x8x4_t;

#ifdef __aarch64__
typedef struct poly64x1x4_t {
  poly64x1_t val[4];
} poly64x1x4_t;

typedef struct poly64x2x4_t {
  poly64x2_t val[4];
} poly64x2x4_t;

#endif

#define __ai static inline __attribute__((__always_inline__, __nodebug__))

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  return __ret;
}
#else
__ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint8x16_t __noswap_vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai uint32x4_t __noswap_vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  return __ret;
}
#else
__ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint16x8_t __noswap_vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  return __ret;
}
#else
__ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int8x16_t __noswap_vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  return __ret;
}
#else
__ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  return __ret;
}
#else
__ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai int32x4_t __noswap_vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  return __ret;
}
#else
__ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int16x8_t __noswap_vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  return __ret;
}
#else
__ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint8x8_t __noswap_vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai uint32x2_t __noswap_vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  return __ret;
}
#else
__ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai uint16x4_t __noswap_vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  return __ret;
}
#else
__ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int8x8_t __noswap_vabd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  return __ret;
}
#else
__ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  return __ret;
}
#else
__ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai int32x2_t __noswap_vabd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  return __ret;
}
#else
__ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai int16x4_t __noswap_vabd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vabsq_s8(int8x16_t __p0) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 32);
  return __ret;
}
#else
__ai int8x16_t vabsq_s8(int8x16_t __p0) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vabsq_f32(float32x4_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 41);
  return __ret;
}
#else
__ai float32x4_t vabsq_f32(float32x4_t __p0) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 41);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vabsq_s32(int32x4_t __p0) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 34);
  return __ret;
}
#else
__ai int32x4_t vabsq_s32(int32x4_t __p0) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vabsq_s16(int16x8_t __p0) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 33);
  return __ret;
}
#else
__ai int16x8_t vabsq_s16(int16x8_t __p0) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vabs_s8(int8x8_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__p0, 0);
  return __ret;
}
#else
__ai int8x8_t vabs_s8(int8x8_t __p0) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vabs_f32(float32x2_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 9);
  return __ret;
}
#else
__ai float32x2_t vabs_f32(float32x2_t __p0) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 9);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vabs_s32(int32x2_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 2);
  return __ret;
}
#else
__ai int32x2_t vabs_s32(int32x2_t __p0) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vabs_s16(int16x4_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__p0, 1);
  return __ret;
}
#else
__ai int16x4_t vabs_s16(int16x4_t __p0) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint64x2_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int64x2_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  float32x2_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = __p0 + __p1;
  return __ret;
}
#else
__ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = __rev0 + __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  return __ret;
}
#else
__ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai uint16x4_t __noswap_vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai uint32x2_t __noswap_vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  return __ret;
}
#else
__ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint8x8_t __noswap_vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  return __ret;
}
#else
__ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai int16x4_t __noswap_vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  return __ret;
}
#else
__ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai int32x2_t __noswap_vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  return __ret;
}
#else
__ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int8x8_t __noswap_vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint64x2_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int64x2_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = __p0 & __p1;
  return __ret;
}
#else
__ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = __rev0 & __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint64x2_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int64x2_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = __p0 & ~__p1;
  return __ret;
}
#else
__ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = __rev0 & ~__rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
  poly8x8_t __ret;
  __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
  return __ret;
}
#else
__ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __ret;
  __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
  poly16x4_t __ret;
  __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 5);
  return __ret;
}
#else
__ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  poly16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  poly16x4_t __ret;
  __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 5);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
  poly8x16_t __ret;
  __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
  return __ret;
}
#else
__ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x16_t __ret;
  __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
  poly16x8_t __ret;
  __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 37);
  return __ret;
}
#else
__ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  poly16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  poly16x8_t __ret;
  __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 37);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
  return __ret;
}
#else
__ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  return __ret;
}
#else
__ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  uint64x2_t __ret;
  __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  return __ret;
}
#else
__ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  uint64x2_t __ret;
  __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 49);
  return __ret;
}
#else
__ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 49);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
  return __ret;
}
#else
__ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  return __ret;
}
#else
__ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  return __ret;
}
#else
__ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  int64x2_t __ret;
  __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 35);
  return __ret;
}
#else
__ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  int64x2_t __ret;
  __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 35);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  return __ret;
}
#else
__ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
  return __ret;
}
#else
__ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
  return __ret;
}
#else
__ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
  uint64x1_t __ret;
  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
  return __ret;
}
#else
__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
  uint64x1_t __ret;
  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 17);
  return __ret;
}
#else
__ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 17);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
  return __ret;
}
#else
__ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  return __ret;
}
#else
__ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  return __ret;
}
#else
__ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
  int64x1_t __ret;
  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
  return __ret;
}
#else
__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
  int64x1_t __ret;
  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
  return __ret;
}
#else
__ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 == __p1);
  return __ret;
}
#else
__ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 == __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 >= __p1);
  return __ret;
}
#else
__ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 >= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 > __p1);
  return __ret;
}
#else
__ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 > __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 <= __p1);
  return __ret;
}
#else
__ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 <= __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vclsq_s8(int8x16_t __p0) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 32);
  return __ret;
}
#else
__ai int8x16_t vclsq_s8(int8x16_t __p0) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vclsq_s32(int32x4_t __p0) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 34);
  return __ret;
}
#else
__ai int32x4_t vclsq_s32(int32x4_t __p0) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vclsq_s16(int16x8_t __p0) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 33);
  return __ret;
}
#else
__ai int16x8_t vclsq_s16(int16x8_t __p0) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vcls_s8(int8x8_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__p0, 0);
  return __ret;
}
#else
__ai int8x8_t vcls_s8(int8x8_t __p0) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vcls_s32(int32x2_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__p0, 2);
  return __ret;
}
#else
__ai int32x2_t vcls_s32(int32x2_t __p0) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vcls_s16(int16x4_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__p0, 1);
  return __ret;
}
#else
__ai int16x4_t vcls_s16(int16x4_t __p0) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0 < __p1);
  return __ret;
}
#else
__ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__rev0 < __rev1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 48);
  return __ret;
}
#else
__ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 48);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 50);
  return __ret;
}
#else
__ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 49);
  return __ret;
}
#else
__ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 49);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vclzq_s8(int8x16_t __p0) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 32);
  return __ret;
}
#else
__ai int8x16_t vclzq_s8(int8x16_t __p0) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vclzq_s32(int32x4_t __p0) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 34);
  return __ret;
}
#else
__ai int32x4_t vclzq_s32(int32x4_t __p0) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vclzq_s16(int16x8_t __p0) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 33);
  return __ret;
}
#else
__ai int16x8_t vclzq_s16(int16x8_t __p0) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vclz_u8(uint8x8_t __p0) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 16);
  return __ret;
}
#else
__ai uint8x8_t vclz_u8(uint8x8_t __p0) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vclz_u32(uint32x2_t __p0) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 18);
  return __ret;
}
#else
__ai uint32x2_t vclz_u32(uint32x2_t __p0) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vclz_u16(uint16x4_t __p0) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 17);
  return __ret;
}
#else
__ai uint16x4_t vclz_u16(uint16x4_t __p0) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 17);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vclz_s8(int8x8_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 0);
  return __ret;
}
#else
__ai int8x8_t vclz_s8(int8x8_t __p0) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vclz_s32(int32x2_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 2);
  return __ret;
}
#else
__ai int32x2_t vclz_s32(int32x2_t __p0) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vclz_s16(int16x4_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 1);
  return __ret;
}
#else
__ai int16x4_t vclz_s16(int16x4_t __p0) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
  poly8x8_t __ret;
  __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 4);
  return __ret;
}
#else
__ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __ret;
  __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 4);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
  poly8x16_t __ret;
  __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 36);
  return __ret;
}
#else
__ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x16_t __ret;
  __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 36);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 48);
  return __ret;
}
#else
__ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 48);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vcntq_s8(int8x16_t __p0) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 32);
  return __ret;
}
#else
__ai int8x16_t vcntq_s8(int8x16_t __p0) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 16);
  return __ret;
}
#else
__ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vcnt_s8(int8x8_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 0);
  return __ret;
}
#else
__ai int8x8_t vcnt_s8(int8x8_t __p0) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
  poly8x16_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#else
__ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x16_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
  poly16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  poly16x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x16_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#else
__ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint8x16_t __noswap_vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x16_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  return __ret;
}
#else
__ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai uint32x4_t __noswap_vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  return __ret;
}
#else
__ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint16x8_t __noswap_vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x16_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#else
__ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int8x16_t __noswap_vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x16_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  return __ret;
}
#else
__ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  float32x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai float32x4_t __noswap_vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
  float32x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
  float16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  float16x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai float16x8_t __noswap_vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
  float16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  return __ret;
}
#else
__ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai int32x4_t __noswap_vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  return __ret;
}
#else
__ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int16x8_t __noswap_vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x8_t vcreate_p8(uint64_t __p0) {
  poly8x8_t __ret;
  __ret = (poly8x8_t)(__p0);
  return __ret;
}
#else
__ai poly8x8_t vcreate_p8(uint64_t __p0) {
  poly8x8_t __ret;
  __ret = (poly8x8_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x4_t vcreate_p16(uint64_t __p0) {
  poly16x4_t __ret;
  __ret = (poly16x4_t)(__p0);
  return __ret;
}
#else
__ai poly16x4_t vcreate_p16(uint64_t __p0) {
  poly16x4_t __ret;
  __ret = (poly16x4_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vcreate_u8(uint64_t __p0) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0);
  return __ret;
}
#else
__ai uint8x8_t vcreate_u8(uint64_t __p0) {
  uint8x8_t __ret;
  __ret = (uint8x8_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcreate_u32(uint64_t __p0) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0);
  return __ret;
}
#else
__ai uint32x2_t vcreate_u32(uint64_t __p0) {
  uint32x2_t __ret;
  __ret = (uint32x2_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vcreate_u64(uint64_t __p0) {
  uint64x1_t __ret;
  __ret = (uint64x1_t)(__p0);
  return __ret;
}
#else
__ai uint64x1_t vcreate_u64(uint64_t __p0) {
  uint64x1_t __ret;
  __ret = (uint64x1_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vcreate_u16(uint64_t __p0) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0);
  return __ret;
}
#else
__ai uint16x4_t vcreate_u16(uint64_t __p0) {
  uint16x4_t __ret;
  __ret = (uint16x4_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vcreate_s8(uint64_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t)(__p0);
  return __ret;
}
#else
__ai int8x8_t vcreate_s8(uint64_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vcreate_f32(uint64_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t)(__p0);
  return __ret;
}
#else
__ai float32x2_t vcreate_f32(uint64_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float16x4_t vcreate_f16(uint64_t __p0) {
  float16x4_t __ret;
  __ret = (float16x4_t)(__p0);
  return __ret;
}
#else
__ai float16x4_t vcreate_f16(uint64_t __p0) {
  float16x4_t __ret;
  __ret = (float16x4_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vcreate_s32(uint64_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t)(__p0);
  return __ret;
}
#else
__ai int32x2_t vcreate_s32(uint64_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vcreate_s64(uint64_t __p0) {
  int64x1_t __ret;
  __ret = (int64x1_t)(__p0);
  return __ret;
}
#else
__ai int64x1_t vcreate_s64(uint64_t __p0) {
  int64x1_t __ret;
  __ret = (int64x1_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vcreate_s16(uint64_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t)(__p0);
  return __ret;
}
#else
__ai int16x4_t vcreate_s16(uint64_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t)(__p0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
  float16x4_t __ret;
  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x16_t)__p0, 8);
  return __ret;
}
#else
__ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float16x4_t __ret;
  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x16_t)__rev0, 8);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai float16x4_t __noswap_vcvt_f16_f32(float32x4_t __p0) {
  float16x4_t __ret;
  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x16_t)__p0, 8);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 50);
  return __ret;
}
#else
__ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 34);
  return __ret;
}
#else
__ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 18);
  return __ret;
}
#else
__ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 2);
  return __ret;
}
#else
__ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 8);
  return __ret;
}
#else
__ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__rev0, 8);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai float32x4_t __noswap_vcvt_f32_f16(float16x4_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 8);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 50); \
  __ret; \
})
#else
#define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 50); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 34); \
  __ret; \
})
#else
#define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 34); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 18); \
  __ret; \
})
#else
#define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 18); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 2); \
  __ret; \
})
#else
#define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 2); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__s0, __p1, 34); \
  __ret; \
})
#else
#define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__rev0, __p1, 34); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__s0, __p1, 2); \
  __ret; \
})
#else
#define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__rev0, __p1, 2); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__s0, __p1, 50); \
  __ret; \
})
#else
#define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__rev0, __p1, 50); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__s0, __p1, 18); \
  __ret; \
})
#else
#define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__rev0, __p1, 18); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__p0, 34);
  return __ret;
}
#else
__ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__rev0, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__p0, 2);
  return __ret;
}
#else
__ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__rev0, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__p0, 50);
  return __ret;
}
#else
__ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__rev0, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__p0, 18);
  return __ret;
}
#else
__ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__rev0, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x8_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x8_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x4_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  poly16x4_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x16_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x16_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x8_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  poly16x8_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x16_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x16_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x4_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint32x4_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x8_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint16x8_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x16_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x16_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x4_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  float32x4_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x4_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int32x4_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x8_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdupq_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int16x8_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x8_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x8_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint32x2_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64x1_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
  __ret; \
})
#else
#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64x1_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x4_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint16x4_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x8_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x8_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  float32x2_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int32x2_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64x1_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
  __ret; \
})
#else
#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64x1_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x4_t __ret; \
  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
  __ret; \
})
#else
#define vdup_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int16x4_t __ret; \
  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x8_t vdup_n_p8(poly8_t __p0) {
  poly8x8_t __ret;
  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai poly8x8_t vdup_n_p8(poly8_t __p0) {
  poly8x8_t __ret;
  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x4_t vdup_n_p16(poly16_t __p0) {
  poly16x4_t __ret;
  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai poly16x4_t vdup_n_p16(poly16_t __p0) {
  poly16x4_t __ret;
  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
  poly8x16_t __ret;
  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
  poly8x16_t __ret;
  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
  poly16x8_t __ret;
  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
  poly16x8_t __ret;
  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
  uint64x2_t __ret;
  __ret = (uint64x2_t) {__p0, __p0};
  return __ret;
}
#else
__ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
  uint64x2_t __ret;
  __ret = (uint64x2_t) {__p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vdupq_n_s8(int8_t __p0) {
  int8x16_t __ret;
  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai int8x16_t vdupq_n_s8(int8_t __p0) {
  int8x16_t __ret;
  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vdupq_n_f32(float32_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai float32x4_t vdupq_n_f32(float32_t __p0) {
  float32x4_t __ret;
  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vdupq_n_f16(__p0) __extension__ ({ \
  float16_t __s0 = __p0; \
  float16x8_t __ret; \
  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
  __ret; \
})
#else
#define vdupq_n_f16(__p0) __extension__ ({ \
  float16_t __s0 = __p0; \
  float16x8_t __ret; \
  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vdupq_n_s32(int32_t __p0) {
  int32x4_t __ret;
  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai int32x4_t vdupq_n_s32(int32_t __p0) {
  int32x4_t __ret;
  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t vdupq_n_s64(int64_t __p0) {
  int64x2_t __ret;
  __ret = (int64x2_t) {__p0, __p0};
  return __ret;
}
#else
__ai int64x2_t vdupq_n_s64(int64_t __p0) {
  int64x2_t __ret;
  __ret = (int64x2_t) {__p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vdupq_n_s16(int16_t __p0) {
  int16x8_t __ret;
  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai int16x8_t vdupq_n_s16(int16_t __p0) {
  int16x8_t __ret;
  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vdup_n_u8(uint8_t __p0) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai uint8x8_t vdup_n_u8(uint8_t __p0) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vdup_n_u32(uint32_t __p0) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) {__p0, __p0};
  return __ret;
}
#else
__ai uint32x2_t vdup_n_u32(uint32_t __p0) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) {__p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
  uint64x1_t __ret;
  __ret = (uint64x1_t) {__p0};
  return __ret;
}
#else
__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
  uint64x1_t __ret;
  __ret = (uint64x1_t) {__p0};
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vdup_n_u16(uint16_t __p0) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai uint16x4_t vdup_n_u16(uint16_t __p0) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vdup_n_s8(int8_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai int8x8_t vdup_n_s8(int8_t __p0) {
  int8x8_t __ret;
  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vdup_n_f32(float32_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t) {__p0, __p0};
  return __ret;
}
#else
__ai float32x2_t vdup_n_f32(float32_t __p0) {
  float32x2_t __ret;
  __ret = (float32x2_t) {__p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vdup_n_f16(__p0) __extension__ ({ \
  float16_t __s0 = __p0; \
  float16x4_t __ret; \
  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
  __ret; \
})
#else
#define vdup_n_f16(__p0) __extension__ ({ \
  float16_t __s0 = __p0; \
  float16x4_t __ret; \
  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vdup_n_s32(int32_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t) {__p0, __p0};
  return __ret;
}
#else
__ai int32x2_t vdup_n_s32(int32_t __p0) {
  int32x2_t __ret;
  __ret = (int32x2_t) {__p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vdup_n_s64(int64_t __p0) {
  int64x1_t __ret;
  __ret = (int64x1_t) {__p0};
  return __ret;
}
#else
__ai int64x1_t vdup_n_s64(int64_t __p0) {
  int64x1_t __ret;
  __ret = (int64x1_t) {__p0};
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vdup_n_s16(int16_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
  return __ret;
}
#else
__ai int16x4_t vdup_n_s16(int16_t __p0) {
  int16x4_t __ret;
  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint64x2_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int64x2_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
  uint64x1_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
  int64x1_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = __p0 ^ __p1;
  return __ret;
}
#else
__ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = __rev0 ^ __rev1;
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x8_t __s1 = __p1; \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
  __ret; \
})
#else
#define vext_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x8_t __s1 = __p1; \
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x4_t __s1 = __p1; \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
  __ret; \
})
#else
#define vext_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x4_t __s1 = __p1; \
  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x16_t __s0 = __p0; \
  poly8x16_t __s1 = __p1; \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
  __ret; \
})
#else
#define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x16_t __s0 = __p0; \
  poly8x16_t __s1 = __p1; \
  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8_t __s0 = __p0; \
  poly16x8_t __s1 = __p1; \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
  __ret; \
})
#else
#define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8_t __s0 = __p0; \
  poly16x8_t __s1 = __p1; \
  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x16_t __s0 = __p0; \
  uint8x16_t __s1 = __p1; \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
  __ret; \
})
#else
#define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x16_t __s0 = __p0; \
  uint8x16_t __s1 = __p1; \
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  uint32x4_t __s1 = __p1; \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
  __ret; \
})
#else
#define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  uint32x4_t __s1 = __p1; \
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x2_t __s0 = __p0; \
  uint64x2_t __s1 = __p1; \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  __ret; \
})
#else
#define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x2_t __s0 = __p0; \
  uint64x2_t __s1 = __p1; \
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x8_t __s0 = __p0; \
  uint16x8_t __s1 = __p1; \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
  __ret; \
})
#else
#define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x8_t __s0 = __p0; \
  uint16x8_t __s1 = __p1; \
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x16_t __s0 = __p0; \
  int8x16_t __s1 = __p1; \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
  __ret; \
})
#else
#define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x16_t __s0 = __p0; \
  int8x16_t __s1 = __p1; \
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32x4_t __s1 = __p1; \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 41); \
  __ret; \
})
#else
#define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32x4_t __s1 = __p1; \
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 41); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  int32x4_t __s1 = __p1; \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  __ret; \
})
#else
#define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  int32x4_t __s1 = __p1; \
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x2_t __s0 = __p0; \
  int64x2_t __s1 = __p1; \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
  __ret; \
})
#else
#define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x2_t __s0 = __p0; \
  int64x2_t __s1 = __p1; \
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x8_t __s0 = __p0; \
  int16x8_t __s1 = __p1; \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  __ret; \
})
#else
#define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x8_t __s0 = __p0; \
  int16x8_t __s1 = __p1; \
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x8_t __s1 = __p1; \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
  __ret; \
})
#else
#define vext_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x8_t __s1 = __p1; \
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __s1 = __p1; \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
  __ret; \
})
#else
#define vext_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __s1 = __p1; \
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64x1_t __s1 = __p1; \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  __ret; \
})
#else
#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64x1_t __s1 = __p1; \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x4_t __s1 = __p1; \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
  __ret; \
})
#else
#define vext_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x4_t __s1 = __p1; \
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x8_t __s1 = __p1; \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
  __ret; \
})
#else
#define vext_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x8_t __s1 = __p1; \
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __s1 = __p1; \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 9); \
  __ret; \
})
#else
#define vext_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __s1 = __p1; \
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 9); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __s1 = __p1; \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  __ret; \
})
#else
#define vext_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __s1 = __p1; \
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64x1_t __s1 = __p1; \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  __ret; \
})
#else
#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64x1_t __s1 = __p1; \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vext_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x4_t __s1 = __p1; \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  __ret; \
})
#else
#define vext_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x4_t __s1 = __p1; \
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  return __ret;
}
#else
__ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai float32x4_t __noswap_vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  float32x4_t __ret;
  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  return __ret;
}
#else
__ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai float32x2_t __noswap_vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  float32x2_t __ret;
  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
  poly8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#else
__ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai poly8x8_t __noswap_vget_high_p8(poly8x16_t __p0) {
  poly8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
  poly16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#else
__ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
  uint8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#else
__ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai uint8x8_t __noswap_vget_high_u8(uint8x16_t __p0) {
  uint8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
  uint32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  return __ret;
}
#else
__ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x2_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai uint32x2_t __noswap_vget_high_u32(uint32x4_t __p0) {
  uint32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
  uint64x1_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 1);
  return __ret;
}
#else
__ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x1_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
  uint16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#else
__ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai uint16x4_t __noswap_vget_high_u16(uint16x8_t __p0) {
  uint16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vget_high_s8(int8x16_t __p0) {
  int8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#else
__ai int8x8_t vget_high_s8(int8x16_t __p0) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
__ai int8x8_t __noswap_vget_high_s8(int8x16_t __p0) {
  int8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vget_high_f32(float32x4_t __p0) {
  float32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  return __ret;
}
#else
__ai float32x2_t vget_high_f32(float32x4_t __p0) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x2_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai float32x2_t __noswap_vget_high_f32(float32x4_t __p0) {
  float32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float16x4_t vget_high_f16(float16x8_t __p0) {
  float16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#else
__ai float16x4_t vget_high_f16(float16x8_t __p0) {
  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  float16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai float16x4_t __noswap_vget_high_f16(float16x8_t __p0) {
  float16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vget_high_s32(int32x4_t __p0) {
  int32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  return __ret;
}
#else
__ai int32x2_t vget_high_s32(int32x4_t __p0) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x2_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
__ai int32x2_t __noswap_vget_high_s32(int32x4_t __p0) {
  int32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vget_high_s64(int64x2_t __p0) {
  int64x1_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 1);
  return __ret;
}
#else
__ai int64x1_t vget_high_s64(int64x2_t __p0) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x1_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vget_high_s16(int16x8_t __p0) {
  int16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#else
__ai int16x4_t vget_high_s16(int16x8_t __p0) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
__ai int16x4_t __noswap_vget_high_s16(int16x8_t __p0) {
  int16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8_t __ret; \
  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8_t __ret; \
  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x8_t __s0 = __p0; \
  poly8_t __ret; \
  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16_t __ret; \
  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  poly16_t __ret; \
  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x4_t __s0 = __p0; \
  poly16_t __ret; \
  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x16_t __s0 = __p0; \
  poly8_t __ret; \
  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x16_t __s0 = __p0; \
  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8_t __ret; \
  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_p8(__p0, __p1) __extension__ ({ \
  poly8x16_t __s0 = __p0; \
  poly8_t __ret; \
  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x8_t __s0 = __p0; \
  poly16_t __ret; \
  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x8_t __s0 = __p0; \
  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly16_t __ret; \
  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_p16(__p0, __p1) __extension__ ({ \
  poly16x8_t __s0 = __p0; \
  poly16_t __ret; \
  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x16_t __s0 = __p0; \
  uint8_t __ret; \
  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x16_t __s0 = __p0; \
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8_t __ret; \
  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x16_t __s0 = __p0; \
  uint8_t __ret; \
  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  uint32_t __ret; \
  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint32_t __ret; \
  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x4_t __s0 = __p0; \
  uint32_t __ret; \
  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x2_t __s0 = __p0; \
  uint64_t __ret; \
  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x2_t __s0 = __p0; \
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint64_t __ret; \
  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x2_t __s0 = __p0; \
  uint64_t __ret; \
  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x8_t __s0 = __p0; \
  uint16_t __ret; \
  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x8_t __s0 = __p0; \
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint16_t __ret; \
  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x8_t __s0 = __p0; \
  uint16_t __ret; \
  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
  int8x16_t __s0 = __p0; \
  int8_t __ret; \
  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
  int8x16_t __s0 = __p0; \
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8_t __ret; \
  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_s8(__p0, __p1) __extension__ ({ \
  int8x16_t __s0 = __p0; \
  int8_t __ret; \
  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32_t __ret; \
  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  float32_t __ret; \
  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_f32(__p0, __p1) __extension__ ({ \
  float32x4_t __s0 = __p0; \
  float32_t __ret; \
  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  int32_t __ret; \
  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int32_t __ret; \
  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_s32(__p0, __p1) __extension__ ({ \
  int32x4_t __s0 = __p0; \
  int32_t __ret; \
  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
  int64x2_t __s0 = __p0; \
  int64_t __ret; \
  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
  int64x2_t __s0 = __p0; \
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int64_t __ret; \
  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_s64(__p0, __p1) __extension__ ({ \
  int64x2_t __s0 = __p0; \
  int64_t __ret; \
  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
  int16x8_t __s0 = __p0; \
  int16_t __ret; \
  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
  __ret; \
})
#else
#define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
  int16x8_t __s0 = __p0; \
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  int16_t __ret; \
  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vgetq_lane_s16(__p0, __p1) __extension__ ({ \
  int16x8_t __s0 = __p0; \
  int16_t __ret; \
  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8_t __ret; \
  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8_t __ret; \
  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_u8(__p0, __p1) __extension__ ({ \
  uint8x8_t __s0 = __p0; \
  uint8_t __ret; \
  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32_t __ret; \
  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  uint32_t __ret; \
  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_u32(__p0, __p1) __extension__ ({ \
  uint32x2_t __s0 = __p0; \
  uint32_t __ret; \
  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64_t __ret; \
  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64_t __ret; \
  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_u64(__p0, __p1) __extension__ ({ \
  uint64x1_t __s0 = __p0; \
  uint64_t __ret; \
  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16_t __ret; \
  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  uint16_t __ret; \
  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_u16(__p0, __p1) __extension__ ({ \
  uint16x4_t __s0 = __p0; \
  uint16_t __ret; \
  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8_t __ret; \
  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8_t __ret; \
  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_s8(__p0, __p1) __extension__ ({ \
  int8x8_t __s0 = __p0; \
  int8_t __ret; \
  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32_t __ret; \
  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  float32_t __ret; \
  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_f32(__p0, __p1) __extension__ ({ \
  float32x2_t __s0 = __p0; \
  float32_t __ret; \
  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32_t __ret; \
  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  int32_t __ret; \
  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_s32(__p0, __p1) __extension__ ({ \
  int32x2_t __s0 = __p0; \
  int32_t __ret; \
  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64_t __ret; \
  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64_t __ret; \
  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_s64(__p0, __p1) __extension__ ({ \
  int64x1_t __s0 = __p0; \
  int64_t __ret; \
  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vget_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16_t __ret; \
  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
  __ret; \
})
#else
#define vget_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  int16_t __ret; \
  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
  __ret; \
})
#define __noswap_vget_lane_s16(__p0, __p1) __extension__ ({ \
  int16x4_t __s0 = __p0; \
  int16_t __ret; \
  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
  poly8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  poly8x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
  poly16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  return __ret;
}
#else
__ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  poly16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
  uint8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
  uint32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
  return __ret;
}
#else
__ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x2_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
  uint64x1_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0);
  return __ret;
}
#else
__ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint64x1_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
  uint16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  return __ret;
}
#else
__ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vget_low_s8(int8x16_t __p0) {
  int8x8_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
  return __ret;
}
#else
__ai int8x8_t vget_low_s8(int8x16_t __p0) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float32x2_t vget_low_f32(float32x4_t __p0) {
  float32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
  return __ret;
}
#else
__ai float32x2_t vget_low_f32(float32x4_t __p0) {
  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  float32x2_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai float16x4_t vget_low_f16(float16x8_t __p0) {
  float16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  return __ret;
}
#else
__ai float16x4_t vget_low_f16(float16x8_t __p0) {
  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  float16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vget_low_s32(int32x4_t __p0) {
  int32x2_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
  return __ret;
}
#else
__ai int32x2_t vget_low_s32(int32x4_t __p0) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x2_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int64x1_t vget_low_s64(int64x2_t __p0) {
  int64x1_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0);
  return __ret;
}
#else
__ai int64x1_t vget_low_s64(int64x2_t __p0) {
  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int64x1_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vget_low_s16(int16x8_t __p0) {
  int16x4_t __ret;
  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  return __ret;
}
#else
__ai int16x4_t vget_low_s16(int16x8_t __p0) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  return __ret;
}
#else
__ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  return __ret;
}
#else
__ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  return __ret;
}
#else
__ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  return __ret;
}
#else
__ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  return __ret;
}
#else
__ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  return __ret;
}
#else
__ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  return __ret;
}
#else
__ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  return __ret;
}
#else
__ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  return __ret;
}
#else
__ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  return __ret;
}
#else
__ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  return __ret;
}
#else
__ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x16_t __ret;
  __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  return __ret;
}
#else
__ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint32x4_t __ret;
  __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  return __ret;
}
#else
__ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint16x8_t __ret;
  __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  return __ret;
}
#else
__ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x16_t __ret;
  __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  return __ret;
}
#else
__ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int32x4_t __ret;
  __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  return __ret;
}
#else
__ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int16x8_t __ret;
  __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  return __ret;
}
#else
__ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  uint8x8_t __ret;
  __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  return __ret;
}
#else
__ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  uint32x2_t __ret;
  __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  return __ret;
}
#else
__ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  uint16x4_t __ret;
  __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  return __ret;
}
#else
__ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  int8x8_t __ret;
  __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  return __ret;
}
#else
__ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  int32x2_t __ret;
  __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
__ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  return __ret;
}
#else
__ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  int16x4_t __ret;
  __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  return __ret;
}
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_p8(__p0) __extension__ ({ \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
  __ret; \
})
#else
#define vld1_p8(__p0) __extension__ ({ \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_p16(__p0) __extension__ ({ \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
  __ret; \
})
#else
#define vld1_p16(__p0) __extension__ ({ \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_p8(__p0) __extension__ ({ \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
  __ret; \
})
#else
#define vld1q_p8(__p0) __extension__ ({ \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_p16(__p0) __extension__ ({ \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
  __ret; \
})
#else
#define vld1q_p16(__p0) __extension__ ({ \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_u8(__p0) __extension__ ({ \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
  __ret; \
})
#else
#define vld1q_u8(__p0) __extension__ ({ \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_u32(__p0) __extension__ ({ \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
  __ret; \
})
#else
#define vld1q_u32(__p0) __extension__ ({ \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_u64(__p0) __extension__ ({ \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
  __ret; \
})
#else
#define vld1q_u64(__p0) __extension__ ({ \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_u16(__p0) __extension__ ({ \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
  __ret; \
})
#else
#define vld1q_u16(__p0) __extension__ ({ \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_s8(__p0) __extension__ ({ \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
  __ret; \
})
#else
#define vld1q_s8(__p0) __extension__ ({ \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_f32(__p0) __extension__ ({ \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
  __ret; \
})
#else
#define vld1q_f32(__p0) __extension__ ({ \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_f16(__p0) __extension__ ({ \
  float16x8_t __ret; \
  __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
  __ret; \
})
#else
#define vld1q_f16(__p0) __extension__ ({ \
  float16x8_t __ret; \
  __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_s32(__p0) __extension__ ({ \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
  __ret; \
})
#else
#define vld1q_s32(__p0) __extension__ ({ \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_s64(__p0) __extension__ ({ \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
  __ret; \
})
#else
#define vld1q_s64(__p0) __extension__ ({ \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_s16(__p0) __extension__ ({ \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
  __ret; \
})
#else
#define vld1q_s16(__p0) __extension__ ({ \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_u8(__p0) __extension__ ({ \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
  __ret; \
})
#else
#define vld1_u8(__p0) __extension__ ({ \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_u32(__p0) __extension__ ({ \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
  __ret; \
})
#else
#define vld1_u32(__p0) __extension__ ({ \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_u64(__p0) __extension__ ({ \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
  __ret; \
})
#else
#define vld1_u64(__p0) __extension__ ({ \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_u16(__p0) __extension__ ({ \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
  __ret; \
})
#else
#define vld1_u16(__p0) __extension__ ({ \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_s8(__p0) __extension__ ({ \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
  __ret; \
})
#else
#define vld1_s8(__p0) __extension__ ({ \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_f32(__p0) __extension__ ({ \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
  __ret; \
})
#else
#define vld1_f32(__p0) __extension__ ({ \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_f16(__p0) __extension__ ({ \
  float16x4_t __ret; \
  __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
  __ret; \
})
#else
#define vld1_f16(__p0) __extension__ ({ \
  float16x4_t __ret; \
  __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_s32(__p0) __extension__ ({ \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
  __ret; \
})
#else
#define vld1_s32(__p0) __extension__ ({ \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_s64(__p0) __extension__ ({ \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
  __ret; \
})
#else
#define vld1_s64(__p0) __extension__ ({ \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_s16(__p0) __extension__ ({ \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
  __ret; \
})
#else
#define vld1_s16(__p0) __extension__ ({ \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_p8(__p0) __extension__ ({ \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
  __ret; \
})
#else
#define vld1_dup_p8(__p0) __extension__ ({ \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_p16(__p0) __extension__ ({ \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
  __ret; \
})
#else
#define vld1_dup_p16(__p0) __extension__ ({ \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_p8(__p0) __extension__ ({ \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
  __ret; \
})
#else
#define vld1q_dup_p8(__p0) __extension__ ({ \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_p16(__p0) __extension__ ({ \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
  __ret; \
})
#else
#define vld1q_dup_p16(__p0) __extension__ ({ \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_u8(__p0) __extension__ ({ \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
  __ret; \
})
#else
#define vld1q_dup_u8(__p0) __extension__ ({ \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_u32(__p0) __extension__ ({ \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
  __ret; \
})
#else
#define vld1q_dup_u32(__p0) __extension__ ({ \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_u64(__p0) __extension__ ({ \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
  __ret; \
})
#else
#define vld1q_dup_u64(__p0) __extension__ ({ \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_u16(__p0) __extension__ ({ \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
  __ret; \
})
#else
#define vld1q_dup_u16(__p0) __extension__ ({ \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_s8(__p0) __extension__ ({ \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
  __ret; \
})
#else
#define vld1q_dup_s8(__p0) __extension__ ({ \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_f32(__p0) __extension__ ({ \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
  __ret; \
})
#else
#define vld1q_dup_f32(__p0) __extension__ ({ \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_f16(__p0) __extension__ ({ \
  float16x8_t __ret; \
  __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
  __ret; \
})
#else
#define vld1q_dup_f16(__p0) __extension__ ({ \
  float16x8_t __ret; \
  __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_s32(__p0) __extension__ ({ \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
  __ret; \
})
#else
#define vld1q_dup_s32(__p0) __extension__ ({ \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_s64(__p0) __extension__ ({ \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
  __ret; \
})
#else
#define vld1q_dup_s64(__p0) __extension__ ({ \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_dup_s16(__p0) __extension__ ({ \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
  __ret; \
})
#else
#define vld1q_dup_s16(__p0) __extension__ ({ \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_u8(__p0) __extension__ ({ \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
  __ret; \
})
#else
#define vld1_dup_u8(__p0) __extension__ ({ \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_u32(__p0) __extension__ ({ \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
  __ret; \
})
#else
#define vld1_dup_u32(__p0) __extension__ ({ \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_u64(__p0) __extension__ ({ \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
  __ret; \
})
#else
#define vld1_dup_u64(__p0) __extension__ ({ \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_u16(__p0) __extension__ ({ \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
  __ret; \
})
#else
#define vld1_dup_u16(__p0) __extension__ ({ \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_s8(__p0) __extension__ ({ \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
  __ret; \
})
#else
#define vld1_dup_s8(__p0) __extension__ ({ \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_f32(__p0) __extension__ ({ \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
  __ret; \
})
#else
#define vld1_dup_f32(__p0) __extension__ ({ \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_f16(__p0) __extension__ ({ \
  float16x4_t __ret; \
  __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
  __ret; \
})
#else
#define vld1_dup_f16(__p0) __extension__ ({ \
  float16x4_t __ret; \
  __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_s32(__p0) __extension__ ({ \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
  __ret; \
})
#else
#define vld1_dup_s32(__p0) __extension__ ({ \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_s64(__p0) __extension__ ({ \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
  __ret; \
})
#else
#define vld1_dup_s64(__p0) __extension__ ({ \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_dup_s16(__p0) __extension__ ({ \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
  __ret; \
})
#else
#define vld1_dup_s16(__p0) __extension__ ({ \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8_t __s1 = __p1; \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
  __ret; \
})
#else
#define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8_t __s1 = __p1; \
  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x8_t __ret; \
  __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4_t __s1 = __p1; \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
  __ret; \
})
#else
#define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4_t __s1 = __p1; \
  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  poly16x4_t __ret; \
  __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x16_t __s1 = __p1; \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
  __ret; \
})
#else
#define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x16_t __s1 = __p1; \
  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x16_t __ret; \
  __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8_t __s1 = __p1; \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
  __ret; \
})
#else
#define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8_t __s1 = __p1; \
  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  poly16x8_t __ret; \
  __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x16_t __s1 = __p1; \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
  __ret; \
})
#else
#define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x16_t __s1 = __p1; \
  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x16_t __ret; \
  __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4_t __s1 = __p1; \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
  __ret; \
})
#else
#define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4_t __s1 = __p1; \
  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  uint32x4_t __ret; \
  __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x2_t __s1 = __p1; \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
  __ret; \
})
#else
#define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x2_t __s1 = __p1; \
  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  uint64x2_t __ret; \
  __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x8_t __s1 = __p1; \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
  __ret; \
})
#else
#define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x8_t __s1 = __p1; \
  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint16x8_t __ret; \
  __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x16_t __s1 = __p1; \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
  __ret; \
})
#else
#define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x16_t __s1 = __p1; \
  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x16_t __ret; \
  __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x4_t __s1 = __p1; \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
  __ret; \
})
#else
#define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x4_t __s1 = __p1; \
  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  float32x4_t __ret; \
  __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x8_t __s1 = __p1; \
  float16x8_t __ret; \
  __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
  __ret; \
})
#else
#define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x8_t __s1 = __p1; \
  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  float16x8_t __ret; \
  __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x4_t __s1 = __p1; \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
  __ret; \
})
#else
#define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x4_t __s1 = __p1; \
  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  int32x4_t __ret; \
  __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x2_t __s1 = __p1; \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
  __ret; \
})
#else
#define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x2_t __s1 = __p1; \
  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  int64x2_t __ret; \
  __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x8_t __s1 = __p1; \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
  __ret; \
})
#else
#define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x8_t __s1 = __p1; \
  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  int16x8_t __ret; \
  __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x8_t __s1 = __p1; \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
  __ret; \
})
#else
#define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x8_t __s1 = __p1; \
  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x8_t __ret; \
  __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x2_t __s1 = __p1; \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
  __ret; \
})
#else
#define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x2_t __s1 = __p1; \
  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  uint32x2_t __ret; \
  __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x1_t __s1 = __p1; \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
  __ret; \
})
#else
#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  uint64x1_t __s1 = __p1; \
  uint64x1_t __ret; \
  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x4_t __s1 = __p1; \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
  __ret; \
})
#else
#define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x4_t __s1 = __p1; \
  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  uint16x4_t __ret; \
  __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x8_t __s1 = __p1; \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
  __ret; \
})
#else
#define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x8_t __s1 = __p1; \
  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x8_t __ret; \
  __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x2_t __s1 = __p1; \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
  __ret; \
})
#else
#define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x2_t __s1 = __p1; \
  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  float32x2_t __ret; \
  __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x4_t __s1 = __p1; \
  float16x4_t __ret; \
  __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
  __ret; \
})
#else
#define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x4_t __s1 = __p1; \
  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  float16x4_t __ret; \
  __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x2_t __s1 = __p1; \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
  __ret; \
})
#else
#define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x2_t __s1 = __p1; \
  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  int32x2_t __ret; \
  __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x1_t __s1 = __p1; \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
  __ret; \
})
#else
#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  int64x1_t __s1 = __p1; \
  int64x1_t __ret; \
  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x4_t __s1 = __p1; \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
  __ret; \
})
#else
#define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x4_t __s1 = __p1; \
  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  int16x4_t __ret; \
  __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_p8(__p0) __extension__ ({ \
  poly8x8x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 4); \
  __ret; \
})
#else
#define vld2_p8(__p0) __extension__ ({ \
  poly8x8x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 4); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_p16(__p0) __extension__ ({ \
  poly16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 5); \
  __ret; \
})
#else
#define vld2_p16(__p0) __extension__ ({ \
  poly16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 5); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_p8(__p0) __extension__ ({ \
  poly8x16x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 36); \
  __ret; \
})
#else
#define vld2q_p8(__p0) __extension__ ({ \
  poly8x16x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 36); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_p16(__p0) __extension__ ({ \
  poly16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 37); \
  __ret; \
})
#else
#define vld2q_p16(__p0) __extension__ ({ \
  poly16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 37); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_u8(__p0) __extension__ ({ \
  uint8x16x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 48); \
  __ret; \
})
#else
#define vld2q_u8(__p0) __extension__ ({ \
  uint8x16x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 48); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_u32(__p0) __extension__ ({ \
  uint32x4x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 50); \
  __ret; \
})
#else
#define vld2q_u32(__p0) __extension__ ({ \
  uint32x4x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 50); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_u16(__p0) __extension__ ({ \
  uint16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 49); \
  __ret; \
})
#else
#define vld2q_u16(__p0) __extension__ ({ \
  uint16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 49); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_s8(__p0) __extension__ ({ \
  int8x16x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 32); \
  __ret; \
})
#else
#define vld2q_s8(__p0) __extension__ ({ \
  int8x16x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 32); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_f32(__p0) __extension__ ({ \
  float32x4x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 41); \
  __ret; \
})
#else
#define vld2q_f32(__p0) __extension__ ({ \
  float32x4x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 41); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_f16(__p0) __extension__ ({ \
  float16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 40); \
  __ret; \
})
#else
#define vld2q_f16(__p0) __extension__ ({ \
  float16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 40); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_s32(__p0) __extension__ ({ \
  int32x4x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 34); \
  __ret; \
})
#else
#define vld2q_s32(__p0) __extension__ ({ \
  int32x4x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 34); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_s16(__p0) __extension__ ({ \
  int16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 33); \
  __ret; \
})
#else
#define vld2q_s16(__p0) __extension__ ({ \
  int16x8x2_t __ret; \
  __builtin_neon_vld2q_v(&__ret, __p0, 33); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_u8(__p0) __extension__ ({ \
  uint8x8x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 16); \
  __ret; \
})
#else
#define vld2_u8(__p0) __extension__ ({ \
  uint8x8x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 16); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_u32(__p0) __extension__ ({ \
  uint32x2x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 18); \
  __ret; \
})
#else
#define vld2_u32(__p0) __extension__ ({ \
  uint32x2x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 18); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_u64(__p0) __extension__ ({ \
  uint64x1x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 19); \
  __ret; \
})
#else
#define vld2_u64(__p0) __extension__ ({ \
  uint64x1x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_u16(__p0) __extension__ ({ \
  uint16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 17); \
  __ret; \
})
#else
#define vld2_u16(__p0) __extension__ ({ \
  uint16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 17); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_s8(__p0) __extension__ ({ \
  int8x8x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 0); \
  __ret; \
})
#else
#define vld2_s8(__p0) __extension__ ({ \
  int8x8x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 0); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_f32(__p0) __extension__ ({ \
  float32x2x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 9); \
  __ret; \
})
#else
#define vld2_f32(__p0) __extension__ ({ \
  float32x2x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 9); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_f16(__p0) __extension__ ({ \
  float16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 8); \
  __ret; \
})
#else
#define vld2_f16(__p0) __extension__ ({ \
  float16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 8); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_s32(__p0) __extension__ ({ \
  int32x2x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 2); \
  __ret; \
})
#else
#define vld2_s32(__p0) __extension__ ({ \
  int32x2x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 2); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_s64(__p0) __extension__ ({ \
  int64x1x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 3); \
  __ret; \
})
#else
#define vld2_s64(__p0) __extension__ ({ \
  int64x1x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_s16(__p0) __extension__ ({ \
  int16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 1); \
  __ret; \
})
#else
#define vld2_s16(__p0) __extension__ ({ \
  int16x4x2_t __ret; \
  __builtin_neon_vld2_v(&__ret, __p0, 1); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_p8(__p0) __extension__ ({ \
  poly8x8x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
  __ret; \
})
#else
#define vld2_dup_p8(__p0) __extension__ ({ \
  poly8x8x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_p16(__p0) __extension__ ({ \
  poly16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
  __ret; \
})
#else
#define vld2_dup_p16(__p0) __extension__ ({ \
  poly16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_u8(__p0) __extension__ ({ \
  uint8x8x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
  __ret; \
})
#else
#define vld2_dup_u8(__p0) __extension__ ({ \
  uint8x8x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_u32(__p0) __extension__ ({ \
  uint32x2x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
  __ret; \
})
#else
#define vld2_dup_u32(__p0) __extension__ ({ \
  uint32x2x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_u64(__p0) __extension__ ({ \
  uint64x1x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
  __ret; \
})
#else
#define vld2_dup_u64(__p0) __extension__ ({ \
  uint64x1x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_u16(__p0) __extension__ ({ \
  uint16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
  __ret; \
})
#else
#define vld2_dup_u16(__p0) __extension__ ({ \
  uint16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_s8(__p0) __extension__ ({ \
  int8x8x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
  __ret; \
})
#else
#define vld2_dup_s8(__p0) __extension__ ({ \
  int8x8x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_f32(__p0) __extension__ ({ \
  float32x2x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
  __ret; \
})
#else
#define vld2_dup_f32(__p0) __extension__ ({ \
  float32x2x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_f16(__p0) __extension__ ({ \
  float16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
  __ret; \
})
#else
#define vld2_dup_f16(__p0) __extension__ ({ \
  float16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_s32(__p0) __extension__ ({ \
  int32x2x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
  __ret; \
})
#else
#define vld2_dup_s32(__p0) __extension__ ({ \
  int32x2x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_s64(__p0) __extension__ ({ \
  int64x1x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
  __ret; \
})
#else
#define vld2_dup_s64(__p0) __extension__ ({ \
  int64x1x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_dup_s16(__p0) __extension__ ({ \
  int16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
  __ret; \
})
#else
#define vld2_dup_s16(__p0) __extension__ ({ \
  int16x4x2_t __ret; \
  __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8x2_t __s1 = __p1; \
  poly8x8x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
  __ret; \
})
#else
#define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8x2_t __s1 = __p1; \
  poly8x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x8x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4x2_t __s1 = __p1; \
  poly16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
  __ret; \
})
#else
#define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4x2_t __s1 = __p1; \
  poly16x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  poly16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8x2_t __s1 = __p1; \
  poly16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
  __ret; \
})
#else
#define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8x2_t __s1 = __p1; \
  poly16x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  poly16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4x2_t __s1 = __p1; \
  uint32x4x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
  __ret; \
})
#else
#define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4x2_t __s1 = __p1; \
  uint32x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  uint32x4x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x8x2_t __s1 = __p1; \
  uint16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
  __ret; \
})
#else
#define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x8x2_t __s1 = __p1; \
  uint16x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  uint16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x4x2_t __s1 = __p1; \
  float32x4x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 41); \
  __ret; \
})
#else
#define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x4x2_t __s1 = __p1; \
  float32x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  float32x4x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 41); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x8x2_t __s1 = __p1; \
  float16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 40); \
  __ret; \
})
#else
#define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x8x2_t __s1 = __p1; \
  float16x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  float16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 40); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x4x2_t __s1 = __p1; \
  int32x4x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 34); \
  __ret; \
})
#else
#define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x4x2_t __s1 = __p1; \
  int32x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  int32x4x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 34); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x8x2_t __s1 = __p1; \
  int16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 33); \
  __ret; \
})
#else
#define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x8x2_t __s1 = __p1; \
  int16x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  int16x8x2_t __ret; \
  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 33); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x8x2_t __s1 = __p1; \
  uint8x8x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
  __ret; \
})
#else
#define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  uint8x8x2_t __s1 = __p1; \
  uint8x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  uint8x8x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x2x2_t __s1 = __p1; \
  uint32x2x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
  __ret; \
})
#else
#define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x2x2_t __s1 = __p1; \
  uint32x2x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  uint32x2x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x4x2_t __s1 = __p1; \
  uint16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
  __ret; \
})
#else
#define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  uint16x4x2_t __s1 = __p1; \
  uint16x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  uint16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x8x2_t __s1 = __p1; \
  int8x8x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
  __ret; \
})
#else
#define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  int8x8x2_t __s1 = __p1; \
  int8x8x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  int8x8x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x2x2_t __s1 = __p1; \
  float32x2x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 9); \
  __ret; \
})
#else
#define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  float32x2x2_t __s1 = __p1; \
  float32x2x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  float32x2x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 9); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x4x2_t __s1 = __p1; \
  float16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 8); \
  __ret; \
})
#else
#define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  float16x4x2_t __s1 = __p1; \
  float16x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  float16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 8); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x2x2_t __s1 = __p1; \
  int32x2x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 2); \
  __ret; \
})
#else
#define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  int32x2x2_t __s1 = __p1; \
  int32x2x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  int32x2x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 2); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x4x2_t __s1 = __p1; \
  int16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 1); \
  __ret; \
})
#else
#define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  int16x4x2_t __s1 = __p1; \
  int16x4x2_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  int16x4x2_t __ret; \
  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 1); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_p8(__p0) __extension__ ({ \
  poly8x8x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 4); \
  __ret; \
})
#else
#define vld3_p8(__p0) __extension__ ({ \
  poly8x8x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 4); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_p16(__p0) __extension__ ({ \
  poly16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 5); \
  __ret; \
})
#else
#define vld3_p16(__p0) __extension__ ({ \
  poly16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 5); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_p8(__p0) __extension__ ({ \
  poly8x16x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 36); \
  __ret; \
})
#else
#define vld3q_p8(__p0) __extension__ ({ \
  poly8x16x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 36); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_p16(__p0) __extension__ ({ \
  poly16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 37); \
  __ret; \
})
#else
#define vld3q_p16(__p0) __extension__ ({ \
  poly16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 37); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_u8(__p0) __extension__ ({ \
  uint8x16x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 48); \
  __ret; \
})
#else
#define vld3q_u8(__p0) __extension__ ({ \
  uint8x16x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 48); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_u32(__p0) __extension__ ({ \
  uint32x4x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 50); \
  __ret; \
})
#else
#define vld3q_u32(__p0) __extension__ ({ \
  uint32x4x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 50); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_u16(__p0) __extension__ ({ \
  uint16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 49); \
  __ret; \
})
#else
#define vld3q_u16(__p0) __extension__ ({ \
  uint16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 49); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_s8(__p0) __extension__ ({ \
  int8x16x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 32); \
  __ret; \
})
#else
#define vld3q_s8(__p0) __extension__ ({ \
  int8x16x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 32); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_f32(__p0) __extension__ ({ \
  float32x4x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 41); \
  __ret; \
})
#else
#define vld3q_f32(__p0) __extension__ ({ \
  float32x4x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 41); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_f16(__p0) __extension__ ({ \
  float16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 40); \
  __ret; \
})
#else
#define vld3q_f16(__p0) __extension__ ({ \
  float16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 40); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_s32(__p0) __extension__ ({ \
  int32x4x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 34); \
  __ret; \
})
#else
#define vld3q_s32(__p0) __extension__ ({ \
  int32x4x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 34); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_s16(__p0) __extension__ ({ \
  int16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 33); \
  __ret; \
})
#else
#define vld3q_s16(__p0) __extension__ ({ \
  int16x8x3_t __ret; \
  __builtin_neon_vld3q_v(&__ret, __p0, 33); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_u8(__p0) __extension__ ({ \
  uint8x8x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 16); \
  __ret; \
})
#else
#define vld3_u8(__p0) __extension__ ({ \
  uint8x8x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 16); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_u32(__p0) __extension__ ({ \
  uint32x2x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 18); \
  __ret; \
})
#else
#define vld3_u32(__p0) __extension__ ({ \
  uint32x2x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 18); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_u64(__p0) __extension__ ({ \
  uint64x1x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 19); \
  __ret; \
})
#else
#define vld3_u64(__p0) __extension__ ({ \
  uint64x1x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_u16(__p0) __extension__ ({ \
  uint16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 17); \
  __ret; \
})
#else
#define vld3_u16(__p0) __extension__ ({ \
  uint16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 17); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_s8(__p0) __extension__ ({ \
  int8x8x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 0); \
  __ret; \
})
#else
#define vld3_s8(__p0) __extension__ ({ \
  int8x8x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 0); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_f32(__p0) __extension__ ({ \
  float32x2x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 9); \
  __ret; \
})
#else
#define vld3_f32(__p0) __extension__ ({ \
  float32x2x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 9); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_f16(__p0) __extension__ ({ \
  float16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 8); \
  __ret; \
})
#else
#define vld3_f16(__p0) __extension__ ({ \
  float16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 8); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_s32(__p0) __extension__ ({ \
  int32x2x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 2); \
  __ret; \
})
#else
#define vld3_s32(__p0) __extension__ ({ \
  int32x2x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 2); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_s64(__p0) __extension__ ({ \
  int64x1x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 3); \
  __ret; \
})
#else
#define vld3_s64(__p0) __extension__ ({ \
  int64x1x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_s16(__p0) __extension__ ({ \
  int16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 1); \
  __ret; \
})
#else
#define vld3_s16(__p0) __extension__ ({ \
  int16x4x3_t __ret; \
  __builtin_neon_vld3_v(&__ret, __p0, 1); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_p8(__p0) __extension__ ({ \
  poly8x8x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
  __ret; \
})
#else
#define vld3_dup_p8(__p0) __extension__ ({ \
  poly8x8x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_p16(__p0) __extension__ ({ \
  poly16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
  __ret; \
})
#else
#define vld3_dup_p16(__p0) __extension__ ({ \
  poly16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_u8(__p0) __extension__ ({ \
  uint8x8x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
  __ret; \
})
#else
#define vld3_dup_u8(__p0) __extension__ ({ \
  uint8x8x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_u32(__p0) __extension__ ({ \
  uint32x2x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
  __ret; \
})
#else
#define vld3_dup_u32(__p0) __extension__ ({ \
  uint32x2x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_u64(__p0) __extension__ ({ \
  uint64x1x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
  __ret; \
})
#else
#define vld3_dup_u64(__p0) __extension__ ({ \
  uint64x1x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_u16(__p0) __extension__ ({ \
  uint16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
  __ret; \
})
#else
#define vld3_dup_u16(__p0) __extension__ ({ \
  uint16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_s8(__p0) __extension__ ({ \
  int8x8x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
  __ret; \
})
#else
#define vld3_dup_s8(__p0) __extension__ ({ \
  int8x8x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_f32(__p0) __extension__ ({ \
  float32x2x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
  __ret; \
})
#else
#define vld3_dup_f32(__p0) __extension__ ({ \
  float32x2x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_f16(__p0) __extension__ ({ \
  float16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
  __ret; \
})
#else
#define vld3_dup_f16(__p0) __extension__ ({ \
  float16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_s32(__p0) __extension__ ({ \
  int32x2x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
  __ret; \
})
#else
#define vld3_dup_s32(__p0) __extension__ ({ \
  int32x2x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_s64(__p0) __extension__ ({ \
  int64x1x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
  __ret; \
})
#else
#define vld3_dup_s64(__p0) __extension__ ({ \
  int64x1x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_dup_s16(__p0) __extension__ ({ \
  int16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
  __ret; \
})
#else
#define vld3_dup_s16(__p0) __extension__ ({ \
  int16x4x3_t __ret; \
  __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8x3_t __s1 = __p1; \
  poly8x8x3_t __ret; \
  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
  __ret; \
})
#else
#define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  poly8x8x3_t __s1 = __p1; \
  poly8x8x3_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  poly8x8x3_t __ret; \
  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4x3_t __s1 = __p1; \
  poly16x4x3_t __ret; \
  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
  __ret; \
})
#else
#define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x4x3_t __s1 = __p1; \
  poly16x4x3_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  poly16x4x3_t __ret; \
  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8x3_t __s1 = __p1; \
  poly16x8x3_t __ret; \
  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
  __ret; \
})
#else
#define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  poly16x8x3_t __s1 = __p1; \
  poly16x8x3_t __rev1; \
  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  poly16x8x3_t __ret; \
  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
 \
  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  __ret; \
})
#endif

#ifdef __LITTLE_ENDIAN__
#define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  uint32x4x3_t __s1 = __p1; \
  uint32x4x3_t __ret; \
  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
  __ret; \
})
#else