Multiply & multiply-accumulate

Multiply, widening and saturating doubling multiply, fused multiply-add/subtract, multiply-accumulate, and polynomial multiply.

__ai __attribute__((target("bf16,neon")))
float32x4_t vbfmlalbq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2)

__ai __attribute__((target("bf16,neon")))
float32x4_t vbfmlaltq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2)

__ai __attribute__((target("fullfp16,neon")))
float16x8_t vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2)

__ai __attribute__((target("fullfp16,neon")))
float16x4_t vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2)

__ai __attribute__((target("fullfp16,neon")))
float16x8_t vfmsq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2)

__ai __attribute__((target("fullfp16,neon")))
float16x4_t vfms_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2)

__ai __attribute__((target("fullfp16,neon")))
float16x8_t vmulq_f16(float16x8_t __p0, float16x8_t __p1)

__ai __attribute__((target("fullfp16,neon")))
float16x4_t vmul_f16(float16x4_t __p0, float16x4_t __p1)

#define vmulq_n_f16(__p0, __p1) __extension__ ({
  float16x8_t __ret;
  float16x8_t __s0 = __p0;
  float16_t __s1 = __p1;
  __ret = __s0 * (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1};
  __ret;
})

#define vmul_n_f16(__p0, __p1) __extension__ ({
  float16x4_t __ret;
  float16x4_t __s0 = __p0;
  float16_t __s1 = __p1;
  __ret = __s0 * (float16x4_t) {__s1, __s1, __s1, __s1};
  __ret;
})

__ai __attribute__((target("neon")))
uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2)

__ai __attribute__((target("neon")))
uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2)

__ai __attribute__((target("neon")))
int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2)

__ai __attribute__((target("neon")))
uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2)

__ai __attribute__((target("neon")))
uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2)

__ai __attribute__((target("neon")))
uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2)

__ai __attribute__((target("neon")))
int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2)

__ai __attribute__((target("neon")))
int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("neon")))
int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vmlaq_lane_u32(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({
  uint32x4_t __ret_54;
  uint32x4_t __s0_54 = __p0_54;
  uint32x4_t __s1_54 = __p1_54;
  uint32x2_t __s2_54 = __p2_54;
  __ret_54 = __s0_54 + __s1_54 * splatq_lane_u32(__s2_54, __p3_54);
  __ret_54;
})

#define vmlaq_lane_u16(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({
  uint16x8_t __ret_56;
  uint16x8_t __s0_56 = __p0_56;
  uint16x8_t __s1_56 = __p1_56;
  uint16x4_t __s2_56 = __p2_56;
  __ret_56 = __s0_56 + __s1_56 * splatq_lane_u16(__s2_56, __p3_56);
  __ret_56;
})

#define vmlaq_lane_f32(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({
  float32x4_t __ret_58;
  float32x4_t __s0_58 = __p0_58;
  float32x4_t __s1_58 = __p1_58;
  float32x2_t __s2_58 = __p2_58;
  __ret_58 = __s0_58 + __s1_58 * splatq_lane_f32(__s2_58, __p3_58);
  __ret_58;
})

#define vmlaq_lane_s32(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({
  int32x4_t __ret_60;
  int32x4_t __s0_60 = __p0_60;
  int32x4_t __s1_60 = __p1_60;
  int32x2_t __s2_60 = __p2_60;
  __ret_60 = __s0_60 + __s1_60 * splatq_lane_s32(__s2_60, __p3_60);
  __ret_60;
})

#define vmlaq_lane_s16(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({
  int16x8_t __ret_62;
  int16x8_t __s0_62 = __p0_62;
  int16x8_t __s1_62 = __p1_62;
  int16x4_t __s2_62 = __p2_62;
  __ret_62 = __s0_62 + __s1_62 * splatq_lane_s16(__s2_62, __p3_62);
  __ret_62;
})

#define vmla_lane_u32(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({
  uint32x2_t __ret_64;
  uint32x2_t __s0_64 = __p0_64;
  uint32x2_t __s1_64 = __p1_64;
  uint32x2_t __s2_64 = __p2_64;
  __ret_64 = __s0_64 + __s1_64 * splat_lane_u32(__s2_64, __p3_64);
  __ret_64;
})

#define vmla_lane_u16(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({
  uint16x4_t __ret_66;
  uint16x4_t __s0_66 = __p0_66;
  uint16x4_t __s1_66 = __p1_66;
  uint16x4_t __s2_66 = __p2_66;
  __ret_66 = __s0_66 + __s1_66 * splat_lane_u16(__s2_66, __p3_66);
  __ret_66;
})

#define vmla_lane_f32(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({
  float32x2_t __ret_68;
  float32x2_t __s0_68 = __p0_68;
  float32x2_t __s1_68 = __p1_68;
  float32x2_t __s2_68 = __p2_68;
  __ret_68 = __s0_68 + __s1_68 * splat_lane_f32(__s2_68, __p3_68);
  __ret_68;
})

#define vmla_lane_s32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({
  int32x2_t __ret_70;
  int32x2_t __s0_70 = __p0_70;
  int32x2_t __s1_70 = __p1_70;
  int32x2_t __s2_70 = __p2_70;
  __ret_70 = __s0_70 + __s1_70 * splat_lane_s32(__s2_70, __p3_70);
  __ret_70;
})

#define vmla_lane_s16(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({
  int16x4_t __ret_72;
  int16x4_t __s0_72 = __p0_72;
  int16x4_t __s1_72 = __p1_72;
  int16x4_t __s2_72 = __p2_72;
  __ret_72 = __s0_72 + __s1_72 * splat_lane_s16(__s2_72, __p3_72);
  __ret_72;
})

__ai __attribute__((target("neon")))
uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2)

__ai __attribute__((target("neon")))
uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2)

__ai __attribute__((target("neon")))
int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2)

__ai __attribute__((target("neon")))
uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2)

__ai __attribute__((target("neon")))
uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2)

__ai __attribute__((target("neon")))
uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2)

__ai __attribute__((target("neon")))
int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2)

__ai __attribute__((target("neon")))
int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("neon")))
int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vmlsq_lane_u32(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({
  uint32x4_t __ret_74;
  uint32x4_t __s0_74 = __p0_74;
  uint32x4_t __s1_74 = __p1_74;
  uint32x2_t __s2_74 = __p2_74;
  __ret_74 = __s0_74 - __s1_74 * splatq_lane_u32(__s2_74, __p3_74);
  __ret_74;
})

#define vmlsq_lane_u16(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({
  uint16x8_t __ret_76;
  uint16x8_t __s0_76 = __p0_76;
  uint16x8_t __s1_76 = __p1_76;
  uint16x4_t __s2_76 = __p2_76;
  __ret_76 = __s0_76 - __s1_76 * splatq_lane_u16(__s2_76, __p3_76);
  __ret_76;
})

#define vmlsq_lane_f32(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({
  float32x4_t __ret_78;
  float32x4_t __s0_78 = __p0_78;
  float32x4_t __s1_78 = __p1_78;
  float32x2_t __s2_78 = __p2_78;
  __ret_78 = __s0_78 - __s1_78 * splatq_lane_f32(__s2_78, __p3_78);
  __ret_78;
})

#define vmlsq_lane_s32(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({
  int32x4_t __ret_80;
  int32x4_t __s0_80 = __p0_80;
  int32x4_t __s1_80 = __p1_80;
  int32x2_t __s2_80 = __p2_80;
  __ret_80 = __s0_80 - __s1_80 * splatq_lane_s32(__s2_80, __p3_80);
  __ret_80;
})

#define vmlsq_lane_s16(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({
  int16x8_t __ret_82;
  int16x8_t __s0_82 = __p0_82;
  int16x8_t __s1_82 = __p1_82;
  int16x4_t __s2_82 = __p2_82;
  __ret_82 = __s0_82 - __s1_82 * splatq_lane_s16(__s2_82, __p3_82);
  __ret_82;
})

#define vmls_lane_u32(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({
  uint32x2_t __ret_84;
  uint32x2_t __s0_84 = __p0_84;
  uint32x2_t __s1_84 = __p1_84;
  uint32x2_t __s2_84 = __p2_84;
  __ret_84 = __s0_84 - __s1_84 * splat_lane_u32(__s2_84, __p3_84);
  __ret_84;
})

#define vmls_lane_u16(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({
  uint16x4_t __ret_86;
  uint16x4_t __s0_86 = __p0_86;
  uint16x4_t __s1_86 = __p1_86;
  uint16x4_t __s2_86 = __p2_86;
  __ret_86 = __s0_86 - __s1_86 * splat_lane_u16(__s2_86, __p3_86);
  __ret_86;
})

#define vmls_lane_f32(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({
  float32x2_t __ret_88;
  float32x2_t __s0_88 = __p0_88;
  float32x2_t __s1_88 = __p1_88;
  float32x2_t __s2_88 = __p2_88;
  __ret_88 = __s0_88 - __s1_88 * splat_lane_f32(__s2_88, __p3_88);
  __ret_88;
})

#define vmls_lane_s32(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({
  int32x2_t __ret_90;
  int32x2_t __s0_90 = __p0_90;
  int32x2_t __s1_90 = __p1_90;
  int32x2_t __s2_90 = __p2_90;
  __ret_90 = __s0_90 - __s1_90 * splat_lane_s32(__s2_90, __p3_90);
  __ret_90;
})

#define vmls_lane_s16(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({
  int16x4_t __ret_92;
  int16x4_t __s0_92 = __p0_92;
  int16x4_t __s1_92 = __p1_92;
  int16x4_t __s2_92 = __p2_92;
  __ret_92 = __s0_92 - __s1_92 * splat_lane_s16(__s2_92, __p3_92);
  __ret_92;
})

__ai __attribute__((target("neon")))
uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1)

__ai __attribute__((target("neon")))
uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1)

__ai __attribute__((target("neon")))
uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1)

__ai __attribute__((target("neon")))
int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1)

__ai __attribute__((target("neon")))
float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1)

__ai __attribute__((target("neon")))
uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1)

__ai __attribute__((target("neon")))
uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1)

__ai __attribute__((target("neon")))
uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1)

__ai __attribute__((target("neon")))
int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1)

__ai __attribute__((target("neon")))
float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1)

__ai __attribute__((target("neon")))
int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1)

__ai __attribute__((target("neon")))
int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1)

__ai __attribute__((target("neon")))
poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1)

__ai __attribute__((target("neon")))
poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1)

#define vmulq_lane_u32(__p0_94, __p1_94, __p2_94) __extension__ ({
  uint32x4_t __ret_94;
  uint32x4_t __s0_94 = __p0_94;
  uint32x2_t __s1_94 = __p1_94;
  __ret_94 = __s0_94 * splatq_lane_u32(__s1_94, __p2_94);
  __ret_94;
})

#define vmulq_lane_u16(__p0_96, __p1_96, __p2_96) __extension__ ({
  uint16x8_t __ret_96;
  uint16x8_t __s0_96 = __p0_96;
  uint16x4_t __s1_96 = __p1_96;
  __ret_96 = __s0_96 * splatq_lane_u16(__s1_96, __p2_96);
  __ret_96;
})

#define vmulq_lane_f32(__p0_98, __p1_98, __p2_98) __extension__ ({
  float32x4_t __ret_98;
  float32x4_t __s0_98 = __p0_98;
  float32x2_t __s1_98 = __p1_98;
  __ret_98 = __s0_98 * splatq_lane_f32(__s1_98, __p2_98);
  __ret_98;
})

#define vmulq_lane_s32(__p0_100, __p1_100, __p2_100) __extension__ ({
  int32x4_t __ret_100;
  int32x4_t __s0_100 = __p0_100;
  int32x2_t __s1_100 = __p1_100;
  __ret_100 = __s0_100 * splatq_lane_s32(__s1_100, __p2_100);
  __ret_100;
})

#define vmulq_lane_s16(__p0_102, __p1_102, __p2_102) __extension__ ({
  int16x8_t __ret_102;
  int16x8_t __s0_102 = __p0_102;
  int16x4_t __s1_102 = __p1_102;
  __ret_102 = __s0_102 * splatq_lane_s16(__s1_102, __p2_102);
  __ret_102;
})

#define vmul_lane_u32(__p0_104, __p1_104, __p2_104) __extension__ ({
  uint32x2_t __ret_104;
  uint32x2_t __s0_104 = __p0_104;
  uint32x2_t __s1_104 = __p1_104;
  __ret_104 = __s0_104 * splat_lane_u32(__s1_104, __p2_104);
  __ret_104;
})

#define vmul_lane_u16(__p0_106, __p1_106, __p2_106) __extension__ ({
  uint16x4_t __ret_106;
  uint16x4_t __s0_106 = __p0_106;
  uint16x4_t __s1_106 = __p1_106;
  __ret_106 = __s0_106 * splat_lane_u16(__s1_106, __p2_106);
  __ret_106;
})

#define vmul_lane_f32(__p0_108, __p1_108, __p2_108) __extension__ ({
  float32x2_t __ret_108;
  float32x2_t __s0_108 = __p0_108;
  float32x2_t __s1_108 = __p1_108;
  __ret_108 = __s0_108 * splat_lane_f32(__s1_108, __p2_108);
  __ret_108;
})

#define vmul_lane_s32(__p0_110, __p1_110, __p2_110) __extension__ ({
  int32x2_t __ret_110;
  int32x2_t __s0_110 = __p0_110;
  int32x2_t __s1_110 = __p1_110;
  __ret_110 = __s0_110 * splat_lane_s32(__s1_110, __p2_110);
  __ret_110;
})

#define vmul_lane_s16(__p0_112, __p1_112, __p2_112) __extension__ ({
  int16x4_t __ret_112;
  int16x4_t __s0_112 = __p0_112;
  int16x4_t __s1_112 = __p1_112;
  __ret_112 = __s0_112 * splat_lane_s16(__s1_112, __p2_112);
  __ret_112;
})

__ai __attribute__((target("neon")))
uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1)

__ai __attribute__((target("neon")))
uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1)

__ai __attribute__((target("neon")))
float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1)

__ai __attribute__((target("neon")))
uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1)

__ai __attribute__((target("neon")))
float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1)

__ai __attribute__((target("neon")))
int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1)

__ai __attribute__((target("neon")))
uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1)

__ai __attribute__((target("neon")))
uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1)

__ai __attribute__((target("neon")))
uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1)

#define vmull_lane_u32(__p0_114, __p1_114, __p2_114) __extension__ ({
  uint64x2_t __ret_114;
  uint32x2_t __s0_114 = __p0_114;
  uint32x2_t __s1_114 = __p1_114;
  __ret_114 = vmull_u32(__s0_114, splat_lane_u32(__s1_114, __p2_114));
  __ret_114;
})

#define vmull_lane_u16(__p0_116, __p1_116, __p2_116) __extension__ ({
  uint32x4_t __ret_116;
  uint16x4_t __s0_116 = __p0_116;
  uint16x4_t __s1_116 = __p1_116;
  __ret_116 = vmull_u16(__s0_116, splat_lane_u16(__s1_116, __p2_116));
  __ret_116;
})

#define vmull_lane_s32(__p0_118, __p1_118, __p2_118) __extension__ ({
  int64x2_t __ret_118;
  int32x2_t __s0_118 = __p0_118;
  int32x2_t __s1_118 = __p1_118;
  __ret_118 = vmull_s32(__s0_118, splat_lane_s32(__s1_118, __p2_118));
  __ret_118;
})

#define vmull_lane_s16(__p0_120, __p1_120, __p2_120) __extension__ ({
  int32x4_t __ret_120;
  int16x4_t __s0_120 = __p0_120;
  int16x4_t __s1_120 = __p1_120;
  __ret_120 = vmull_s16(__s0_120, splat_lane_s16(__s1_120, __p2_120));
  __ret_120;
})

__ai __attribute__((target("neon")))
uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1)

__ai __attribute__((target("neon")))
uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vqdmlal_lane_s32(__p0_122, __p1_122, __p2_122, __p3_122) __extension__ ({
  int64x2_t __ret_122;
  int64x2_t __s0_122 = __p0_122;
  int32x2_t __s1_122 = __p1_122;
  int32x2_t __s2_122 = __p2_122;
  __ret_122 = vqdmlal_s32(__s0_122, __s1_122, splat_lane_s32(__s2_122, __p3_122));
  __ret_122;
})

#define vqdmlal_lane_s16(__p0_124, __p1_124, __p2_124, __p3_124) __extension__ ({
  int32x4_t __ret_124;
  int32x4_t __s0_124 = __p0_124;
  int16x4_t __s1_124 = __p1_124;
  int16x4_t __s2_124 = __p2_124;
  __ret_124 = vqdmlal_s16(__s0_124, __s1_124, splat_lane_s16(__s2_124, __p3_124));
  __ret_124;
})

__ai __attribute__((target("neon")))
int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vqdmlsl_lane_s32(__p0_126, __p1_126, __p2_126, __p3_126) __extension__ ({
  int64x2_t __ret_126;
  int64x2_t __s0_126 = __p0_126;
  int32x2_t __s1_126 = __p1_126;
  int32x2_t __s2_126 = __p2_126;
  __ret_126 = vqdmlsl_s32(__s0_126, __s1_126, splat_lane_s32(__s2_126, __p3_126));
  __ret_126;
})

#define vqdmlsl_lane_s16(__p0_128, __p1_128, __p2_128, __p3_128) __extension__ ({
  int32x4_t __ret_128;
  int32x4_t __s0_128 = __p0_128;
  int16x4_t __s1_128 = __p1_128;
  int16x4_t __s2_128 = __p2_128;
  __ret_128 = vqdmlsl_s16(__s0_128, __s1_128, splat_lane_s16(__s2_128, __p3_128));
  __ret_128;
})

__ai __attribute__((target("neon")))
int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1)

__ai __attribute__((target("neon")))
int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1)

__ai __attribute__((target("neon")))
int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1)

#define vqdmull_lane_s32(__p0_130, __p1_130, __p2_130) __extension__ ({
  int64x2_t __ret_130;
  int32x2_t __s0_130 = __p0_130;
  int32x2_t __s1_130 = __p1_130;
  __ret_130 = vqdmull_s32(__s0_130, splat_lane_s32(__s1_130, __p2_130));
  __ret_130;
})

#define vqdmull_lane_s16(__p0_132, __p1_132, __p2_132) __extension__ ({
  int32x4_t __ret_132;
  int16x4_t __s0_132 = __p0_132;
  int16x4_t __s1_132 = __p1_132;
  __ret_132 = vqdmull_s16(__s0_132, splat_lane_s16(__s1_132, __p2_132));
  __ret_132;
})

__ai __attribute__((target("neon")))
int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1)

__ai __attribute__((target("neon")))
int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1)

__ai __attribute__((target("neon")))
int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1)

__ai __attribute__((target("v8.1a,neon")))
int32x4_t vqrdmlahq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int16x8_t vqrdmlahq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int32x2_t vqrdmlah_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int16x4_t vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vqrdmlahq_lane_s32(__p0_134, __p1_134, __p2_134, __p3_134) __extension__ ({
  int32x4_t __ret_134;
  int32x4_t __s0_134 = __p0_134;
  int32x4_t __s1_134 = __p1_134;
  int32x2_t __s2_134 = __p2_134;
  __ret_134 = vqrdmlahq_s32(__s0_134, __s1_134, splatq_lane_s32(__s2_134, __p3_134));
  __ret_134;
})

#define vqrdmlahq_lane_s16(__p0_136, __p1_136, __p2_136, __p3_136) __extension__ ({
  int16x8_t __ret_136;
  int16x8_t __s0_136 = __p0_136;
  int16x8_t __s1_136 = __p1_136;
  int16x4_t __s2_136 = __p2_136;
  __ret_136 = vqrdmlahq_s16(__s0_136, __s1_136, splatq_lane_s16(__s2_136, __p3_136));
  __ret_136;
})

#define vqrdmlah_lane_s32(__p0_138, __p1_138, __p2_138, __p3_138) __extension__ ({
  int32x2_t __ret_138;
  int32x2_t __s0_138 = __p0_138;
  int32x2_t __s1_138 = __p1_138;
  int32x2_t __s2_138 = __p2_138;
  __ret_138 = vqrdmlah_s32(__s0_138, __s1_138, splat_lane_s32(__s2_138, __p3_138));
  __ret_138;
})

#define vqrdmlah_lane_s16(__p0_140, __p1_140, __p2_140, __p3_140) __extension__ ({
  int16x4_t __ret_140;
  int16x4_t __s0_140 = __p0_140;
  int16x4_t __s1_140 = __p1_140;
  int16x4_t __s2_140 = __p2_140;
  __ret_140 = vqrdmlah_s16(__s0_140, __s1_140, splat_lane_s16(__s2_140, __p3_140));
  __ret_140;
})

__ai __attribute__((target("v8.1a,neon")))
int32x4_t vqrdmlshq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int16x8_t vqrdmlshq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int32x2_t vqrdmlsh_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int16x4_t vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vqrdmlshq_lane_s32(__p0_142, __p1_142, __p2_142, __p3_142) __extension__ ({
  int32x4_t __ret_142;
  int32x4_t __s0_142 = __p0_142;
  int32x4_t __s1_142 = __p1_142;
  int32x2_t __s2_142 = __p2_142;
  __ret_142 = vqrdmlshq_s32(__s0_142, __s1_142, splatq_lane_s32(__s2_142, __p3_142));
  __ret_142;
})

#define vqrdmlshq_lane_s16(__p0_144, __p1_144, __p2_144, __p3_144) __extension__ ({
  int16x8_t __ret_144;
  int16x8_t __s0_144 = __p0_144;
  int16x8_t __s1_144 = __p1_144;
  int16x4_t __s2_144 = __p2_144;
  __ret_144 = vqrdmlshq_s16(__s0_144, __s1_144, splatq_lane_s16(__s2_144, __p3_144));
  __ret_144;
})

#define vqrdmlsh_lane_s32(__p0_146, __p1_146, __p2_146, __p3_146) __extension__ ({
  int32x2_t __ret_146;
  int32x2_t __s0_146 = __p0_146;
  int32x2_t __s1_146 = __p1_146;
  int32x2_t __s2_146 = __p2_146;
  __ret_146 = vqrdmlsh_s32(__s0_146, __s1_146, splat_lane_s32(__s2_146, __p3_146));
  __ret_146;
})

#define vqrdmlsh_lane_s16(__p0_148, __p1_148, __p2_148, __p3_148) __extension__ ({
  int16x4_t __ret_148;
  int16x4_t __s0_148 = __p0_148;
  int16x4_t __s1_148 = __p1_148;
  int16x4_t __s2_148 = __p2_148;
  __ret_148 = vqrdmlsh_s16(__s0_148, __s1_148, splat_lane_s16(__s2_148, __p3_148));
  __ret_148;
})

__ai __attribute__((target("neon")))
float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2)

__ai __attribute__((target("aes,neon")))
poly128_t vmull_p64(poly64_t __p0, poly64_t __p1)

__ai __attribute__((target("fp16fml,neon")))
float32x4_t vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x2_t vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x4_t vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x2_t vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x4_t vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x2_t vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x4_t vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2)

__ai __attribute__((target("fp16fml,neon")))
float32x2_t vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2)

#define vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({
  float16_t __ret;
  float16_t __s0 = __p0;
  float16_t __s1 = __p1;
  float16x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(float16_t, __builtin_neon_vfmah_lane_f16(__s0, __s1, __s2, __p3));
  __ret;
})

#define vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({
  float16x8_t __ret;
  float16x8_t __s0 = __p0;
  float16x8_t __s1 = __p1;
  float16x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(float16x8_t, __builtin_neon_vfmaq_lane_f16(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __builtin_bit_cast(int8x8_t, __s2), __p3, 40));
  __ret;
})

#define vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({
  float16x4_t __ret;
  float16x4_t __s0 = __p0;
  float16x4_t __s1 = __p1;
  float16x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(float16x4_t, __builtin_neon_vfma_lane_f16(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __builtin_bit_cast(int8x8_t, __s2), __p3, 8));
  __ret;
})

#define vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({
  float16_t __ret;
  float16_t __s0 = __p0;
  float16_t __s1 = __p1;
  float16x8_t __s2 = __p2;
  __ret = __builtin_bit_cast(float16_t, __builtin_neon_vfmah_laneq_f16(__s0, __s1, __s2, __p3));
  __ret;
})

#define vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({
  float16x8_t __ret;
  float16x8_t __s0 = __p0;
  float16x8_t __s1 = __p1;
  float16x8_t __s2 = __p2;
  __ret = __builtin_bit_cast(float16x8_t, __builtin_neon_vfmaq_laneq_f16(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __builtin_bit_cast(int8x16_t, __s2), __p3, 40));
  __ret;
})

#define vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({
  float16x4_t __ret;
  float16x4_t __s0 = __p0;
  float16x4_t __s1 = __p1;
  float16x8_t __s2 = __p2;
  __ret = __builtin_bit_cast(float16x4_t, __builtin_neon_vfma_laneq_f16(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __builtin_bit_cast(int8x16_t, __s2), __p3, 8));
  __ret;
})

#define vfmaq_n_f16(__p0, __p1, __p2) __extension__ ({
  float16x8_t __ret;
  float16x8_t __s0 = __p0;
  float16x8_t __s1 = __p1;
  float16_t __s2 = __p2;
  __ret = vfmaq_f16(__s0, __s1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2});
  __ret;
})

#define vfma_n_f16(__p0, __p1, __p2) __extension__ ({
  float16x4_t __ret;
  float16x4_t __s0 = __p0;
  float16x4_t __s1 = __p1;
  float16_t __s2 = __p2;
  __ret = vfma_f16(__s0, __s1, (float16x4_t) {__s2, __s2, __s2, __s2});
  __ret;
})

#define vfmsh_lane_f16(__p0_250, __p1_250, __p2_250, __p3_250) __extension__ ({
  float16_t __ret_250;
  float16_t __s0_250 = __p0_250;
  float16_t __s1_250 = __p1_250;
  float16x4_t __s2_250 = __p2_250;
  __ret_250 = vfmah_lane_f16(__s0_250, -__s1_250, __s2_250, __p3_250);
  __ret_250;
})

#define vfmsq_lane_f16(__p0_252, __p1_252, __p2_252, __p3_252) __extension__ ({
  float16x8_t __ret_252;
  float16x8_t __s0_252 = __p0_252;
  float16x8_t __s1_252 = __p1_252;
  float16x4_t __s2_252 = __p2_252;
  __ret_252 = vfmaq_lane_f16(__s0_252, -__s1_252, __s2_252, __p3_252);
  __ret_252;
})

#define vfms_lane_f16(__p0_254, __p1_254, __p2_254, __p3_254) __extension__ ({
  float16x4_t __ret_254;
  float16x4_t __s0_254 = __p0_254;
  float16x4_t __s1_254 = __p1_254;
  float16x4_t __s2_254 = __p2_254;
  __ret_254 = vfma_lane_f16(__s0_254, -__s1_254, __s2_254, __p3_254);
  __ret_254;
})

#define vfmsh_laneq_f16(__p0_256, __p1_256, __p2_256, __p3_256) __extension__ ({
  float16_t __ret_256;
  float16_t __s0_256 = __p0_256;
  float16_t __s1_256 = __p1_256;
  float16x8_t __s2_256 = __p2_256;
  __ret_256 = vfmah_laneq_f16(__s0_256, -__s1_256, __s2_256, __p3_256);
  __ret_256;
})

#define vfmsq_laneq_f16(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({
  float16x8_t __ret_258;
  float16x8_t __s0_258 = __p0_258;
  float16x8_t __s1_258 = __p1_258;
  float16x8_t __s2_258 = __p2_258;
  __ret_258 = vfmaq_laneq_f16(__s0_258, -__s1_258, __s2_258, __p3_258);
  __ret_258;
})

#define vfms_laneq_f16(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({
  float16x4_t __ret_260;
  float16x4_t __s0_260 = __p0_260;
  float16x4_t __s1_260 = __p1_260;
  float16x8_t __s2_260 = __p2_260;
  __ret_260 = vfma_laneq_f16(__s0_260, -__s1_260, __s2_260, __p3_260);
  __ret_260;
})

#define vfmsq_n_f16(__p0, __p1, __p2) __extension__ ({
  float16x8_t __ret;
  float16x8_t __s0 = __p0;
  float16x8_t __s1 = __p1;
  float16_t __s2 = __p2;
  __ret = vfmaq_f16(__s0, -__s1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2});
  __ret;
})

#define vfms_n_f16(__p0, __p1, __p2) __extension__ ({
  float16x4_t __ret;
  float16x4_t __s0 = __p0;
  float16x4_t __s1 = __p1;
  float16_t __s2 = __p2;
  __ret = vfma_f16(__s0, -__s1, (float16x4_t) {__s2, __s2, __s2, __s2});
  __ret;
})

#define vmulq_laneq_f16(__p0_262, __p1_262, __p2_262) __extension__ ({
  float16x8_t __ret_262;
  float16x8_t __s0_262 = __p0_262;
  float16x8_t __s1_262 = __p1_262;
  __ret_262 = __s0_262 * splatq_laneq_f16(__s1_262, __p2_262);
  __ret_262;
})

#define vmul_laneq_f16(__p0_264, __p1_264, __p2_264) __extension__ ({
  float16x4_t __ret_264;
  float16x4_t __s0_264 = __p0_264;
  float16x8_t __s1_264 = __p1_264;
  __ret_264 = __s0_264 * splat_laneq_f16(__s1_264, __p2_264);
  __ret_264;
})

__ai __attribute__((target("fullfp16,neon")))
float16x8_t vmulxq_f16(float16x8_t __p0, float16x8_t __p1)

__ai __attribute__((target("fullfp16,neon")))
float16x4_t vmulx_f16(float16x4_t __p0, float16x4_t __p1)

#define vmulxh_lane_f16(__p0, __p1, __p2) __extension__ ({
  float16_t __ret;
  float16_t __s0 = __p0;
  float16x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(float16_t, __builtin_neon_vmulxh_lane_f16(__s0, __s1, __p2));
  __ret;
})

#define vmulxq_lane_f16(__p0_266, __p1_266, __p2_266) __extension__ ({
  float16x8_t __ret_266;
  float16x8_t __s0_266 = __p0_266;
  float16x4_t __s1_266 = __p1_266;
  __ret_266 = vmulxq_f16(__s0_266, splatq_lane_f16(__s1_266, __p2_266));
  __ret_266;
})

#define vmulx_lane_f16(__p0_268, __p1_268, __p2_268) __extension__ ({
  float16x4_t __ret_268;
  float16x4_t __s0_268 = __p0_268;
  float16x4_t __s1_268 = __p1_268;
  __ret_268 = vmulx_f16(__s0_268, splat_lane_f16(__s1_268, __p2_268));
  __ret_268;
})

#define vmulxh_laneq_f16(__p0, __p1, __p2) __extension__ ({
  float16_t __ret;
  float16_t __s0 = __p0;
  float16x8_t __s1 = __p1;
  __ret = __builtin_bit_cast(float16_t, __builtin_neon_vmulxh_laneq_f16(__s0, __s1, __p2));
  __ret;
})

#define vmulxq_laneq_f16(__p0_270, __p1_270, __p2_270) __extension__ ({
  float16x8_t __ret_270;
  float16x8_t __s0_270 = __p0_270;
  float16x8_t __s1_270 = __p1_270;
  __ret_270 = vmulxq_f16(__s0_270, splatq_laneq_f16(__s1_270, __p2_270));
  __ret_270;
})

#define vmulx_laneq_f16(__p0_272, __p1_272, __p2_272) __extension__ ({
  float16x4_t __ret_272;
  float16x4_t __s0_272 = __p0_272;
  float16x8_t __s1_272 = __p1_272;
  __ret_272 = vmulx_f16(__s0_272, splat_laneq_f16(__s1_272, __p2_272));
  __ret_272;
})

#define vmulxq_n_f16(__p0, __p1) __extension__ ({
  float16x8_t __ret;
  float16x8_t __s0 = __p0;
  float16_t __s1 = __p1;
  __ret = vmulxq_f16(__s0, (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1});
  __ret;
})

#define vmulx_n_f16(__p0, __p1) __extension__ ({
  float16x4_t __ret;
  float16x4_t __s0 = __p0;
  float16_t __s1 = __p1;
  __ret = vmulx_f16(__s0, (float16x4_t) {__s1, __s1, __s1, __s1});
  __ret;
})

__ai __attribute__((target("neon")))
float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2)

__ai __attribute__((target("neon")))
float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2)

#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({
  float64_t __ret;
  float64_t __s0 = __p0;
  float64_t __s1 = __p1;
  float64x1_t __s2 = __p2;
  __ret = __builtin_bit_cast(float64_t, __builtin_neon_vfmad_lane_f64(__s0, __s1, __s2, __p3));
  __ret;
})

#define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({
  float32_t __ret;
  float32_t __s0 = __p0;
  float32_t __s1 = __p1;
  float32x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(float32_t, __builtin_neon_vfmas_lane_f32(__s0, __s1, __s2, __p3));
  __ret;
})

#define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({
  float64x2_t __ret;
  float64x2_t __s0 = __p0;
  float64x2_t __s1 = __p1;
  float64x1_t __s2 = __p2;
  __ret = __builtin_bit_cast(float64x2_t, __builtin_neon_vfmaq_lane_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __builtin_bit_cast(int8x8_t, __s2), __p3, 42));
  __ret;
})

#define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({
  float32x4_t __ret;
  float32x4_t __s0 = __p0;
  float32x4_t __s1 = __p1;
  float32x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(float32x4_t, __builtin_neon_vfmaq_lane_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __builtin_bit_cast(int8x8_t, __s2), __p3, 41));
  __ret;
})

#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({
  float64x1_t __ret;
  float64x1_t __s0 = __p0;
  float64x1_t __s1 = __p1;
  float64x1_t __s2 = __p2;
  __ret = __builtin_bit_cast(float64x1_t, __builtin_neon_vfma_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __builtin_bit_cast(int8x8_t, __s2), __p3, 10));
  __ret;
})

#define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({
  float32x2_t __ret;
  float32x2_t __s0 = __p0;
  float32x2_t __s1 = __p1;
  float32x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(float32x2_t, __builtin_neon_vfma_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __builtin_bit_cast(int8x8_t, __s2), __p3, 9));
  __ret;
})

#define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({
  float64_t __ret;
  float64_t __s0 = __p0;
  float64_t __s1 = __p1;
  float64x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(float64_t, __builtin_neon_vfmad_laneq_f64(__s0, __s1, __s2, __p3));
  __ret;
})

#define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({
  float32_t __ret;
  float32_t __s0 = __p0;
  float32_t __s1 = __p1;
  float32x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(float32_t, __builtin_neon_vfmas_laneq_f32(__s0, __s1, __s2, __p3));
  __ret;
})

#define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({
  float64x2_t __ret;
  float64x2_t __s0 = __p0;
  float64x2_t __s1 = __p1;
  float64x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(float64x2_t, __builtin_neon_vfmaq_laneq_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __builtin_bit_cast(int8x16_t, __s2), __p3, 42));
  __ret;
})

#define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({
  float32x4_t __ret;
  float32x4_t __s0 = __p0;
  float32x4_t __s1 = __p1;
  float32x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(float32x4_t, __builtin_neon_vfmaq_laneq_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __builtin_bit_cast(int8x16_t, __s2), __p3, 41));
  __ret;
})

#define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({
  float64x1_t __ret;
  float64x1_t __s0 = __p0;
  float64x1_t __s1 = __p1;
  float64x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(float64x1_t, __builtin_neon_vfma_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __builtin_bit_cast(int8x16_t, __s2), __p3, 10));
  __ret;
})

#define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({
  float32x2_t __ret;
  float32x2_t __s0 = __p0;
  float32x2_t __s1 = __p1;
  float32x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(float32x2_t, __builtin_neon_vfma_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __builtin_bit_cast(int8x16_t, __s2), __p3, 9));
  __ret;
})

__ai __attribute__((target("neon")))
float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2)

__ai __attribute__((target("neon")))
float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2)

__ai __attribute__((target("neon")))
float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2)

__ai __attribute__((target("neon")))
float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2)

#define vfmsd_lane_f64(__p0_442, __p1_442, __p2_442, __p3_442) __extension__ ({
  float64_t __ret_442;
  float64_t __s0_442 = __p0_442;
  float64_t __s1_442 = __p1_442;
  float64x1_t __s2_442 = __p2_442;
  __ret_442 = vfmad_lane_f64(__s0_442, -__s1_442, __s2_442, __p3_442);
  __ret_442;
})

#define vfmss_lane_f32(__p0_443, __p1_443, __p2_443, __p3_443) __extension__ ({
  float32_t __ret_443;
  float32_t __s0_443 = __p0_443;
  float32_t __s1_443 = __p1_443;
  float32x2_t __s2_443 = __p2_443;
  __ret_443 = vfmas_lane_f32(__s0_443, -__s1_443, __s2_443, __p3_443);
  __ret_443;
})

#define vfmsq_lane_f64(__p0_445, __p1_445, __p2_445, __p3_445) __extension__ ({
  float64x2_t __ret_445;
  float64x2_t __s0_445 = __p0_445;
  float64x2_t __s1_445 = __p1_445;
  float64x1_t __s2_445 = __p2_445;
  __ret_445 = vfmaq_lane_f64(__s0_445, -__s1_445, __s2_445, __p3_445);
  __ret_445;
})

#define vfmsq_lane_f32(__p0_447, __p1_447, __p2_447, __p3_447) __extension__ ({
  float32x4_t __ret_447;
  float32x4_t __s0_447 = __p0_447;
  float32x4_t __s1_447 = __p1_447;
  float32x2_t __s2_447 = __p2_447;
  __ret_447 = vfmaq_lane_f32(__s0_447, -__s1_447, __s2_447, __p3_447);
  __ret_447;
})

#define vfms_lane_f64(__p0_449, __p1_449, __p2_449, __p3_449) __extension__ ({
  float64x1_t __ret_449;
  float64x1_t __s0_449 = __p0_449;
  float64x1_t __s1_449 = __p1_449;
  float64x1_t __s2_449 = __p2_449;
  __ret_449 = vfma_lane_f64(__s0_449, -__s1_449, __s2_449, __p3_449);
  __ret_449;
})

#define vfms_lane_f32(__p0_450, __p1_450, __p2_450, __p3_450) __extension__ ({
  float32x2_t __ret_450;
  float32x2_t __s0_450 = __p0_450;
  float32x2_t __s1_450 = __p1_450;
  float32x2_t __s2_450 = __p2_450;
  __ret_450 = vfma_lane_f32(__s0_450, -__s1_450, __s2_450, __p3_450);
  __ret_450;
})

#define vfmsd_laneq_f64(__p0_452, __p1_452, __p2_452, __p3_452) __extension__ ({
  float64_t __ret_452;
  float64_t __s0_452 = __p0_452;
  float64_t __s1_452 = __p1_452;
  float64x2_t __s2_452 = __p2_452;
  __ret_452 = vfmad_laneq_f64(__s0_452, -__s1_452, __s2_452, __p3_452);
  __ret_452;
})

#define vfmss_laneq_f32(__p0_454, __p1_454, __p2_454, __p3_454) __extension__ ({
  float32_t __ret_454;
  float32_t __s0_454 = __p0_454;
  float32_t __s1_454 = __p1_454;
  float32x4_t __s2_454 = __p2_454;
  __ret_454 = vfmas_laneq_f32(__s0_454, -__s1_454, __s2_454, __p3_454);
  __ret_454;
})

#define vfmsq_laneq_f64(__p0_456, __p1_456, __p2_456, __p3_456) __extension__ ({
  float64x2_t __ret_456;
  float64x2_t __s0_456 = __p0_456;
  float64x2_t __s1_456 = __p1_456;
  float64x2_t __s2_456 = __p2_456;
  __ret_456 = vfmaq_laneq_f64(__s0_456, -__s1_456, __s2_456, __p3_456);
  __ret_456;
})

#define vfmsq_laneq_f32(__p0_458, __p1_458, __p2_458, __p3_458) __extension__ ({
  float32x4_t __ret_458;
  float32x4_t __s0_458 = __p0_458;
  float32x4_t __s1_458 = __p1_458;
  float32x4_t __s2_458 = __p2_458;
  __ret_458 = vfmaq_laneq_f32(__s0_458, -__s1_458, __s2_458, __p3_458);
  __ret_458;
})

#define vfms_laneq_f64(__p0_460, __p1_460, __p2_460, __p3_460) __extension__ ({
  float64x1_t __ret_460;
  float64x1_t __s0_460 = __p0_460;
  float64x1_t __s1_460 = __p1_460;
  float64x2_t __s2_460 = __p2_460;
  __ret_460 = vfma_laneq_f64(__s0_460, -__s1_460, __s2_460, __p3_460);
  __ret_460;
})

#define vfms_laneq_f32(__p0_462, __p1_462, __p2_462, __p3_462) __extension__ ({
  float32x2_t __ret_462;
  float32x2_t __s0_462 = __p0_462;
  float32x2_t __s1_462 = __p1_462;
  float32x4_t __s2_462 = __p2_462;
  __ret_462 = vfma_laneq_f32(__s0_462, -__s1_462, __s2_462, __p3_462);
  __ret_462;
})

__ai __attribute__((target("neon")))
float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2)

__ai __attribute__((target("neon")))
float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2)

__ai __attribute__((target("neon")))
float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2)

__ai __attribute__((target("neon")))
float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2)

__ai __attribute__((target("neon")))
float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2)

#define vmlaq_laneq_u32(__p0_464, __p1_464, __p2_464, __p3_464) __extension__ ({
  uint32x4_t __ret_464;
  uint32x4_t __s0_464 = __p0_464;
  uint32x4_t __s1_464 = __p1_464;
  uint32x4_t __s2_464 = __p2_464;
  __ret_464 = __s0_464 + __s1_464 * splatq_laneq_u32(__s2_464, __p3_464);
  __ret_464;
})

#define vmlaq_laneq_u16(__p0_466, __p1_466, __p2_466, __p3_466) __extension__ ({
  uint16x8_t __ret_466;
  uint16x8_t __s0_466 = __p0_466;
  uint16x8_t __s1_466 = __p1_466;
  uint16x8_t __s2_466 = __p2_466;
  __ret_466 = __s0_466 + __s1_466 * splatq_laneq_u16(__s2_466, __p3_466);
  __ret_466;
})

#define vmlaq_laneq_f32(__p0_468, __p1_468, __p2_468, __p3_468) __extension__ ({
  float32x4_t __ret_468;
  float32x4_t __s0_468 = __p0_468;
  float32x4_t __s1_468 = __p1_468;
  float32x4_t __s2_468 = __p2_468;
  __ret_468 = __s0_468 + __s1_468 * splatq_laneq_f32(__s2_468, __p3_468);
  __ret_468;
})

#define vmlaq_laneq_s32(__p0_470, __p1_470, __p2_470, __p3_470) __extension__ ({
  int32x4_t __ret_470;
  int32x4_t __s0_470 = __p0_470;
  int32x4_t __s1_470 = __p1_470;
  int32x4_t __s2_470 = __p2_470;
  __ret_470 = __s0_470 + __s1_470 * splatq_laneq_s32(__s2_470, __p3_470);
  __ret_470;
})

#define vmlaq_laneq_s16(__p0_472, __p1_472, __p2_472, __p3_472) __extension__ ({
  int16x8_t __ret_472;
  int16x8_t __s0_472 = __p0_472;
  int16x8_t __s1_472 = __p1_472;
  int16x8_t __s2_472 = __p2_472;
  __ret_472 = __s0_472 + __s1_472 * splatq_laneq_s16(__s2_472, __p3_472);
  __ret_472;
})

#define vmla_laneq_u32(__p0_474, __p1_474, __p2_474, __p3_474) __extension__ ({
  uint32x2_t __ret_474;
  uint32x2_t __s0_474 = __p0_474;
  uint32x2_t __s1_474 = __p1_474;
  uint32x4_t __s2_474 = __p2_474;
  __ret_474 = __s0_474 + __s1_474 * splat_laneq_u32(__s2_474, __p3_474);
  __ret_474;
})

#define vmla_laneq_u16(__p0_476, __p1_476, __p2_476, __p3_476) __extension__ ({
  uint16x4_t __ret_476;
  uint16x4_t __s0_476 = __p0_476;
  uint16x4_t __s1_476 = __p1_476;
  uint16x8_t __s2_476 = __p2_476;
  __ret_476 = __s0_476 + __s1_476 * splat_laneq_u16(__s2_476, __p3_476);
  __ret_476;
})

#define vmla_laneq_f32(__p0_478, __p1_478, __p2_478, __p3_478) __extension__ ({
  float32x2_t __ret_478;
  float32x2_t __s0_478 = __p0_478;
  float32x2_t __s1_478 = __p1_478;
  float32x4_t __s2_478 = __p2_478;
  __ret_478 = __s0_478 + __s1_478 * splat_laneq_f32(__s2_478, __p3_478);
  __ret_478;
})

#define vmla_laneq_s32(__p0_480, __p1_480, __p2_480, __p3_480) __extension__ ({
  int32x2_t __ret_480;
  int32x2_t __s0_480 = __p0_480;
  int32x2_t __s1_480 = __p1_480;
  int32x4_t __s2_480 = __p2_480;
  __ret_480 = __s0_480 + __s1_480 * splat_laneq_s32(__s2_480, __p3_480);
  __ret_480;
})

#define vmla_laneq_s16(__p0_482, __p1_482, __p2_482, __p3_482) __extension__ ({
  int16x4_t __ret_482;
  int16x4_t __s0_482 = __p0_482;
  int16x4_t __s1_482 = __p1_482;
  int16x8_t __s2_482 = __p2_482;
  __ret_482 = __s0_482 + __s1_482 * splat_laneq_s16(__s2_482, __p3_482);
  __ret_482;
})

#define vmlal_high_lane_u32(__p0_484, __p1_484, __p2_484, __p3_484) __extension__ ({
  uint64x2_t __ret_484;
  uint64x2_t __s0_484 = __p0_484;
  uint32x4_t __s1_484 = __p1_484;
  uint32x2_t __s2_484 = __p2_484;
  __ret_484 = __s0_484 + vmull_u32(vget_high_u32(__s1_484), splat_lane_u32(__s2_484, __p3_484));
  __ret_484;
})

#define vmlal_high_lane_u16(__p0_486, __p1_486, __p2_486, __p3_486) __extension__ ({
  uint32x4_t __ret_486;
  uint32x4_t __s0_486 = __p0_486;
  uint16x8_t __s1_486 = __p1_486;
  uint16x4_t __s2_486 = __p2_486;
  __ret_486 = __s0_486 + vmull_u16(vget_high_u16(__s1_486), splat_lane_u16(__s2_486, __p3_486));
  __ret_486;
})

#define vmlal_high_lane_s32(__p0_488, __p1_488, __p2_488, __p3_488) __extension__ ({
  int64x2_t __ret_488;
  int64x2_t __s0_488 = __p0_488;
  int32x4_t __s1_488 = __p1_488;
  int32x2_t __s2_488 = __p2_488;
  __ret_488 = __s0_488 + vmull_s32(vget_high_s32(__s1_488), splat_lane_s32(__s2_488, __p3_488));
  __ret_488;
})

#define vmlal_high_lane_s16(__p0_490, __p1_490, __p2_490, __p3_490) __extension__ ({
  int32x4_t __ret_490;
  int32x4_t __s0_490 = __p0_490;
  int16x8_t __s1_490 = __p1_490;
  int16x4_t __s2_490 = __p2_490;
  __ret_490 = __s0_490 + vmull_s16(vget_high_s16(__s1_490), splat_lane_s16(__s2_490, __p3_490));
  __ret_490;
})

#define vmlal_high_laneq_u32(__p0_492, __p1_492, __p2_492, __p3_492) __extension__ ({
  uint64x2_t __ret_492;
  uint64x2_t __s0_492 = __p0_492;
  uint32x4_t __s1_492 = __p1_492;
  uint32x4_t __s2_492 = __p2_492;
  __ret_492 = __s0_492 + vmull_u32(vget_high_u32(__s1_492), splat_laneq_u32(__s2_492, __p3_492));
  __ret_492;
})

#define vmlal_high_laneq_u16(__p0_494, __p1_494, __p2_494, __p3_494) __extension__ ({
  uint32x4_t __ret_494;
  uint32x4_t __s0_494 = __p0_494;
  uint16x8_t __s1_494 = __p1_494;
  uint16x8_t __s2_494 = __p2_494;
  __ret_494 = __s0_494 + vmull_u16(vget_high_u16(__s1_494), splat_laneq_u16(__s2_494, __p3_494));
  __ret_494;
})

#define vmlal_high_laneq_s32(__p0_496, __p1_496, __p2_496, __p3_496) __extension__ ({
  int64x2_t __ret_496;
  int64x2_t __s0_496 = __p0_496;
  int32x4_t __s1_496 = __p1_496;
  int32x4_t __s2_496 = __p2_496;
  __ret_496 = __s0_496 + vmull_s32(vget_high_s32(__s1_496), splat_laneq_s32(__s2_496, __p3_496));
  __ret_496;
})

#define vmlal_high_laneq_s16(__p0_498, __p1_498, __p2_498, __p3_498) __extension__ ({
  int32x4_t __ret_498;
  int32x4_t __s0_498 = __p0_498;
  int16x8_t __s1_498 = __p1_498;
  int16x8_t __s2_498 = __p2_498;
  __ret_498 = __s0_498 + vmull_s16(vget_high_s16(__s1_498), splat_laneq_s16(__s2_498, __p3_498));
  __ret_498;
})

#define vmlal_laneq_u32(__p0_500, __p1_500, __p2_500, __p3_500) __extension__ ({
  uint64x2_t __ret_500;
  uint64x2_t __s0_500 = __p0_500;
  uint32x2_t __s1_500 = __p1_500;
  uint32x4_t __s2_500 = __p2_500;
  __ret_500 = __s0_500 + vmull_u32(__s1_500, splat_laneq_u32(__s2_500, __p3_500));
  __ret_500;
})

#define vmlal_laneq_u16(__p0_502, __p1_502, __p2_502, __p3_502) __extension__ ({
  uint32x4_t __ret_502;
  uint32x4_t __s0_502 = __p0_502;
  uint16x4_t __s1_502 = __p1_502;
  uint16x8_t __s2_502 = __p2_502;
  __ret_502 = __s0_502 + vmull_u16(__s1_502, splat_laneq_u16(__s2_502, __p3_502));
  __ret_502;
})

#define vmlal_laneq_s32(__p0_504, __p1_504, __p2_504, __p3_504) __extension__ ({
  int64x2_t __ret_504;
  int64x2_t __s0_504 = __p0_504;
  int32x2_t __s1_504 = __p1_504;
  int32x4_t __s2_504 = __p2_504;
  __ret_504 = __s0_504 + vmull_s32(__s1_504, splat_laneq_s32(__s2_504, __p3_504));
  __ret_504;
})

#define vmlal_laneq_s16(__p0_506, __p1_506, __p2_506, __p3_506) __extension__ ({
  int32x4_t __ret_506;
  int32x4_t __s0_506 = __p0_506;
  int16x4_t __s1_506 = __p1_506;
  int16x8_t __s2_506 = __p2_506;
  __ret_506 = __s0_506 + vmull_s16(__s1_506, splat_laneq_s16(__s2_506, __p3_506));
  __ret_506;
})

__ai __attribute__((target("neon")))
float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2)

__ai __attribute__((target("neon")))
float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2)

#define vmlsq_laneq_u32(__p0_508, __p1_508, __p2_508, __p3_508) __extension__ ({
  uint32x4_t __ret_508;
  uint32x4_t __s0_508 = __p0_508;
  uint32x4_t __s1_508 = __p1_508;
  uint32x4_t __s2_508 = __p2_508;
  __ret_508 = __s0_508 - __s1_508 * splatq_laneq_u32(__s2_508, __p3_508);
  __ret_508;
})

#define vmlsq_laneq_u16(__p0_510, __p1_510, __p2_510, __p3_510) __extension__ ({
  uint16x8_t __ret_510;
  uint16x8_t __s0_510 = __p0_510;
  uint16x8_t __s1_510 = __p1_510;
  uint16x8_t __s2_510 = __p2_510;
  __ret_510 = __s0_510 - __s1_510 * splatq_laneq_u16(__s2_510, __p3_510);
  __ret_510;
})

#define vmlsq_laneq_f32(__p0_512, __p1_512, __p2_512, __p3_512) __extension__ ({
  float32x4_t __ret_512;
  float32x4_t __s0_512 = __p0_512;
  float32x4_t __s1_512 = __p1_512;
  float32x4_t __s2_512 = __p2_512;
  __ret_512 = __s0_512 - __s1_512 * splatq_laneq_f32(__s2_512, __p3_512);
  __ret_512;
})

#define vmlsq_laneq_s32(__p0_514, __p1_514, __p2_514, __p3_514) __extension__ ({
  int32x4_t __ret_514;
  int32x4_t __s0_514 = __p0_514;
  int32x4_t __s1_514 = __p1_514;
  int32x4_t __s2_514 = __p2_514;
  __ret_514 = __s0_514 - __s1_514 * splatq_laneq_s32(__s2_514, __p3_514);
  __ret_514;
})

#define vmlsq_laneq_s16(__p0_516, __p1_516, __p2_516, __p3_516) __extension__ ({
  int16x8_t __ret_516;
  int16x8_t __s0_516 = __p0_516;
  int16x8_t __s1_516 = __p1_516;
  int16x8_t __s2_516 = __p2_516;
  __ret_516 = __s0_516 - __s1_516 * splatq_laneq_s16(__s2_516, __p3_516);
  __ret_516;
})

#define vmls_laneq_u32(__p0_518, __p1_518, __p2_518, __p3_518) __extension__ ({
  uint32x2_t __ret_518;
  uint32x2_t __s0_518 = __p0_518;
  uint32x2_t __s1_518 = __p1_518;
  uint32x4_t __s2_518 = __p2_518;
  __ret_518 = __s0_518 - __s1_518 * splat_laneq_u32(__s2_518, __p3_518);
  __ret_518;
})

#define vmls_laneq_u16(__p0_520, __p1_520, __p2_520, __p3_520) __extension__ ({
  uint16x4_t __ret_520;
  uint16x4_t __s0_520 = __p0_520;
  uint16x4_t __s1_520 = __p1_520;
  uint16x8_t __s2_520 = __p2_520;
  __ret_520 = __s0_520 - __s1_520 * splat_laneq_u16(__s2_520, __p3_520);
  __ret_520;
})

#define vmls_laneq_f32(__p0_522, __p1_522, __p2_522, __p3_522) __extension__ ({
  float32x2_t __ret_522;
  float32x2_t __s0_522 = __p0_522;
  float32x2_t __s1_522 = __p1_522;
  float32x4_t __s2_522 = __p2_522;
  __ret_522 = __s0_522 - __s1_522 * splat_laneq_f32(__s2_522, __p3_522);
  __ret_522;
})

#define vmls_laneq_s32(__p0_524, __p1_524, __p2_524, __p3_524) __extension__ ({
  int32x2_t __ret_524;
  int32x2_t __s0_524 = __p0_524;
  int32x2_t __s1_524 = __p1_524;
  int32x4_t __s2_524 = __p2_524;
  __ret_524 = __s0_524 - __s1_524 * splat_laneq_s32(__s2_524, __p3_524);
  __ret_524;
})

#define vmls_laneq_s16(__p0_526, __p1_526, __p2_526, __p3_526) __extension__ ({
  int16x4_t __ret_526;
  int16x4_t __s0_526 = __p0_526;
  int16x4_t __s1_526 = __p1_526;
  int16x8_t __s2_526 = __p2_526;
  __ret_526 = __s0_526 - __s1_526 * splat_laneq_s16(__s2_526, __p3_526);
  __ret_526;
})

#define vmlsl_high_lane_u32(__p0_528, __p1_528, __p2_528, __p3_528) __extension__ ({
  uint64x2_t __ret_528;
  uint64x2_t __s0_528 = __p0_528;
  uint32x4_t __s1_528 = __p1_528;
  uint32x2_t __s2_528 = __p2_528;
  __ret_528 = __s0_528 - vmull_u32(vget_high_u32(__s1_528), splat_lane_u32(__s2_528, __p3_528));
  __ret_528;
})

#define vmlsl_high_lane_u16(__p0_530, __p1_530, __p2_530, __p3_530) __extension__ ({
  uint32x4_t __ret_530;
  uint32x4_t __s0_530 = __p0_530;
  uint16x8_t __s1_530 = __p1_530;
  uint16x4_t __s2_530 = __p2_530;
  __ret_530 = __s0_530 - vmull_u16(vget_high_u16(__s1_530), splat_lane_u16(__s2_530, __p3_530));
  __ret_530;
})

#define vmlsl_high_lane_s32(__p0_532, __p1_532, __p2_532, __p3_532) __extension__ ({
  int64x2_t __ret_532;
  int64x2_t __s0_532 = __p0_532;
  int32x4_t __s1_532 = __p1_532;
  int32x2_t __s2_532 = __p2_532;
  __ret_532 = __s0_532 - vmull_s32(vget_high_s32(__s1_532), splat_lane_s32(__s2_532, __p3_532));
  __ret_532;
})

#define vmlsl_high_lane_s16(__p0_534, __p1_534, __p2_534, __p3_534) __extension__ ({
  int32x4_t __ret_534;
  int32x4_t __s0_534 = __p0_534;
  int16x8_t __s1_534 = __p1_534;
  int16x4_t __s2_534 = __p2_534;
  __ret_534 = __s0_534 - vmull_s16(vget_high_s16(__s1_534), splat_lane_s16(__s2_534, __p3_534));
  __ret_534;
})

#define vmlsl_high_laneq_u32(__p0_536, __p1_536, __p2_536, __p3_536) __extension__ ({
  uint64x2_t __ret_536;
  uint64x2_t __s0_536 = __p0_536;
  uint32x4_t __s1_536 = __p1_536;
  uint32x4_t __s2_536 = __p2_536;
  __ret_536 = __s0_536 - vmull_u32(vget_high_u32(__s1_536), splat_laneq_u32(__s2_536, __p3_536));
  __ret_536;
})

#define vmlsl_high_laneq_u16(__p0_538, __p1_538, __p2_538, __p3_538) __extension__ ({
  uint32x4_t __ret_538;
  uint32x4_t __s0_538 = __p0_538;
  uint16x8_t __s1_538 = __p1_538;
  uint16x8_t __s2_538 = __p2_538;
  __ret_538 = __s0_538 - vmull_u16(vget_high_u16(__s1_538), splat_laneq_u16(__s2_538, __p3_538));
  __ret_538;
})

#define vmlsl_high_laneq_s32(__p0_540, __p1_540, __p2_540, __p3_540) __extension__ ({
  int64x2_t __ret_540;
  int64x2_t __s0_540 = __p0_540;
  int32x4_t __s1_540 = __p1_540;
  int32x4_t __s2_540 = __p2_540;
  __ret_540 = __s0_540 - vmull_s32(vget_high_s32(__s1_540), splat_laneq_s32(__s2_540, __p3_540));
  __ret_540;
})

#define vmlsl_high_laneq_s16(__p0_542, __p1_542, __p2_542, __p3_542) __extension__ ({
  int32x4_t __ret_542;
  int32x4_t __s0_542 = __p0_542;
  int16x8_t __s1_542 = __p1_542;
  int16x8_t __s2_542 = __p2_542;
  __ret_542 = __s0_542 - vmull_s16(vget_high_s16(__s1_542), splat_laneq_s16(__s2_542, __p3_542));
  __ret_542;
})

#define vmlsl_laneq_u32(__p0_544, __p1_544, __p2_544, __p3_544) __extension__ ({
  uint64x2_t __ret_544;
  uint64x2_t __s0_544 = __p0_544;
  uint32x2_t __s1_544 = __p1_544;
  uint32x4_t __s2_544 = __p2_544;
  __ret_544 = __s0_544 - vmull_u32(__s1_544, splat_laneq_u32(__s2_544, __p3_544));
  __ret_544;
})

#define vmlsl_laneq_u16(__p0_546, __p1_546, __p2_546, __p3_546) __extension__ ({
  uint32x4_t __ret_546;
  uint32x4_t __s0_546 = __p0_546;
  uint16x4_t __s1_546 = __p1_546;
  uint16x8_t __s2_546 = __p2_546;
  __ret_546 = __s0_546 - vmull_u16(__s1_546, splat_laneq_u16(__s2_546, __p3_546));
  __ret_546;
})

#define vmlsl_laneq_s32(__p0_548, __p1_548, __p2_548, __p3_548) __extension__ ({
  int64x2_t __ret_548;
  int64x2_t __s0_548 = __p0_548;
  int32x2_t __s1_548 = __p1_548;
  int32x4_t __s2_548 = __p2_548;
  __ret_548 = __s0_548 - vmull_s32(__s1_548, splat_laneq_s32(__s2_548, __p3_548));
  __ret_548;
})

#define vmlsl_laneq_s16(__p0_550, __p1_550, __p2_550, __p3_550) __extension__ ({
  int32x4_t __ret_550;
  int32x4_t __s0_550 = __p0_550;
  int16x4_t __s1_550 = __p1_550;
  int16x8_t __s2_550 = __p2_550;
  __ret_550 = __s0_550 - vmull_s16(__s1_550, splat_laneq_s16(__s2_550, __p3_550));
  __ret_550;
})

__ai __attribute__((target("neon")))
float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1)

__ai __attribute__((target("neon")))
float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1)

#define vmuld_lane_f64(__p0_570, __p1_570, __p2_570) __extension__ ({
  float64_t __ret_570;
  float64_t __s0_570 = __p0_570;
  float64x1_t __s1_570 = __p1_570;
  __ret_570 = __s0_570 * vget_lane_f64(__s1_570, __p2_570);
  __ret_570;
})

#define vmuls_lane_f32(__p0_571, __p1_571, __p2_571) __extension__ ({
  float32_t __ret_571;
  float32_t __s0_571 = __p0_571;
  float32x2_t __s1_571 = __p1_571;
  __ret_571 = __s0_571 * vget_lane_f32(__s1_571, __p2_571);
  __ret_571;
})

#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({
  float64x1_t __ret;
  float64x1_t __s0 = __p0;
  float64x1_t __s1 = __p1;
  __ret = __builtin_bit_cast(float64x1_t, __builtin_neon_vmul_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 10));
  __ret;
})

#define vmulq_lane_f64(__p0_573, __p1_573, __p2_573) __extension__ ({
  float64x2_t __ret_573;
  float64x2_t __s0_573 = __p0_573;
  float64x1_t __s1_573 = __p1_573;
  __ret_573 = __s0_573 * splatq_lane_f64(__s1_573, __p2_573);
  __ret_573;
})

#define vmuld_laneq_f64(__p0_575, __p1_575, __p2_575) __extension__ ({
  float64_t __ret_575;
  float64_t __s0_575 = __p0_575;
  float64x2_t __s1_575 = __p1_575;
  __ret_575 = __s0_575 * vgetq_lane_f64(__s1_575, __p2_575);
  __ret_575;
})

#define vmuls_laneq_f32(__p0_577, __p1_577, __p2_577) __extension__ ({
  float32_t __ret_577;
  float32_t __s0_577 = __p0_577;
  float32x4_t __s1_577 = __p1_577;
  __ret_577 = __s0_577 * vgetq_lane_f32(__s1_577, __p2_577);
  __ret_577;
})

#define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({
  float64x1_t __ret;
  float64x1_t __s0 = __p0;
  float64x2_t __s1 = __p1;
  __ret = __builtin_bit_cast(float64x1_t, __builtin_neon_vmul_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 10));
  __ret;
})

#define vmulq_laneq_u32(__p0_579, __p1_579, __p2_579) __extension__ ({
  uint32x4_t __ret_579;
  uint32x4_t __s0_579 = __p0_579;
  uint32x4_t __s1_579 = __p1_579;
  __ret_579 = __s0_579 * splatq_laneq_u32(__s1_579, __p2_579);
  __ret_579;
})

#define vmulq_laneq_u16(__p0_581, __p1_581, __p2_581) __extension__ ({
  uint16x8_t __ret_581;
  uint16x8_t __s0_581 = __p0_581;
  uint16x8_t __s1_581 = __p1_581;
  __ret_581 = __s0_581 * splatq_laneq_u16(__s1_581, __p2_581);
  __ret_581;
})

#define vmulq_laneq_f64(__p0_583, __p1_583, __p2_583) __extension__ ({
  float64x2_t __ret_583;
  float64x2_t __s0_583 = __p0_583;
  float64x2_t __s1_583 = __p1_583;
  __ret_583 = __s0_583 * splatq_laneq_f64(__s1_583, __p2_583);
  __ret_583;
})

#define vmulq_laneq_f32(__p0_585, __p1_585, __p2_585) __extension__ ({
  float32x4_t __ret_585;
  float32x4_t __s0_585 = __p0_585;
  float32x4_t __s1_585 = __p1_585;
  __ret_585 = __s0_585 * splatq_laneq_f32(__s1_585, __p2_585);
  __ret_585;
})

#define vmulq_laneq_s32(__p0_587, __p1_587, __p2_587) __extension__ ({
  int32x4_t __ret_587;
  int32x4_t __s0_587 = __p0_587;
  int32x4_t __s1_587 = __p1_587;
  __ret_587 = __s0_587 * splatq_laneq_s32(__s1_587, __p2_587);
  __ret_587;
})

#define vmulq_laneq_s16(__p0_589, __p1_589, __p2_589) __extension__ ({
  int16x8_t __ret_589;
  int16x8_t __s0_589 = __p0_589;
  int16x8_t __s1_589 = __p1_589;
  __ret_589 = __s0_589 * splatq_laneq_s16(__s1_589, __p2_589);
  __ret_589;
})

#define vmul_laneq_u32(__p0_591, __p1_591, __p2_591) __extension__ ({
  uint32x2_t __ret_591;
  uint32x2_t __s0_591 = __p0_591;
  uint32x4_t __s1_591 = __p1_591;
  __ret_591 = __s0_591 * splat_laneq_u32(__s1_591, __p2_591);
  __ret_591;
})

#define vmul_laneq_u16(__p0_593, __p1_593, __p2_593) __extension__ ({
  uint16x4_t __ret_593;
  uint16x4_t __s0_593 = __p0_593;
  uint16x8_t __s1_593 = __p1_593;
  __ret_593 = __s0_593 * splat_laneq_u16(__s1_593, __p2_593);
  __ret_593;
})

#define vmul_laneq_f32(__p0_595, __p1_595, __p2_595) __extension__ ({
  float32x2_t __ret_595;
  float32x2_t __s0_595 = __p0_595;
  float32x4_t __s1_595 = __p1_595;
  __ret_595 = __s0_595 * splat_laneq_f32(__s1_595, __p2_595);
  __ret_595;
})

#define vmul_laneq_s32(__p0_597, __p1_597, __p2_597) __extension__ ({
  int32x2_t __ret_597;
  int32x2_t __s0_597 = __p0_597;
  int32x4_t __s1_597 = __p1_597;
  __ret_597 = __s0_597 * splat_laneq_s32(__s1_597, __p2_597);
  __ret_597;
})

#define vmul_laneq_s16(__p0_599, __p1_599, __p2_599) __extension__ ({
  int16x4_t __ret_599;
  int16x4_t __s0_599 = __p0_599;
  int16x8_t __s1_599 = __p1_599;
  __ret_599 = __s0_599 * splat_laneq_s16(__s1_599, __p2_599);
  __ret_599;
})

__ai __attribute__((target("neon")))
float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1)

__ai __attribute__((target("neon")))
float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1)

__ai __attribute__((target("neon")))
poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1)

__ai __attribute__((target("neon")))
uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1)

__ai __attribute__((target("neon")))
uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1)

__ai __attribute__((target("neon")))
uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1)

__ai __attribute__((target("neon")))
int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1)

#define vmull_high_lane_u32(__p0_601, __p1_601, __p2_601) __extension__ ({
  uint64x2_t __ret_601;
  uint32x4_t __s0_601 = __p0_601;
  uint32x2_t __s1_601 = __p1_601;
  __ret_601 = vmull_u32(vget_high_u32(__s0_601), splat_lane_u32(__s1_601, __p2_601));
  __ret_601;
})

#define vmull_high_lane_u16(__p0_603, __p1_603, __p2_603) __extension__ ({
  uint32x4_t __ret_603;
  uint16x8_t __s0_603 = __p0_603;
  uint16x4_t __s1_603 = __p1_603;
  __ret_603 = vmull_u16(vget_high_u16(__s0_603), splat_lane_u16(__s1_603, __p2_603));
  __ret_603;
})

#define vmull_high_lane_s32(__p0_605, __p1_605, __p2_605) __extension__ ({
  int64x2_t __ret_605;
  int32x4_t __s0_605 = __p0_605;
  int32x2_t __s1_605 = __p1_605;
  __ret_605 = vmull_s32(vget_high_s32(__s0_605), splat_lane_s32(__s1_605, __p2_605));
  __ret_605;
})

#define vmull_high_lane_s16(__p0_607, __p1_607, __p2_607) __extension__ ({
  int32x4_t __ret_607;
  int16x8_t __s0_607 = __p0_607;
  int16x4_t __s1_607 = __p1_607;
  __ret_607 = vmull_s16(vget_high_s16(__s0_607), splat_lane_s16(__s1_607, __p2_607));
  __ret_607;
})

#define vmull_high_laneq_u32(__p0_609, __p1_609, __p2_609) __extension__ ({
  uint64x2_t __ret_609;
  uint32x4_t __s0_609 = __p0_609;
  uint32x4_t __s1_609 = __p1_609;
  __ret_609 = vmull_u32(vget_high_u32(__s0_609), splat_laneq_u32(__s1_609, __p2_609));
  __ret_609;
})

#define vmull_high_laneq_u16(__p0_611, __p1_611, __p2_611) __extension__ ({
  uint32x4_t __ret_611;
  uint16x8_t __s0_611 = __p0_611;
  uint16x8_t __s1_611 = __p1_611;
  __ret_611 = vmull_u16(vget_high_u16(__s0_611), splat_laneq_u16(__s1_611, __p2_611));
  __ret_611;
})

#define vmull_high_laneq_s32(__p0_613, __p1_613, __p2_613) __extension__ ({
  int64x2_t __ret_613;
  int32x4_t __s0_613 = __p0_613;
  int32x4_t __s1_613 = __p1_613;
  __ret_613 = vmull_s32(vget_high_s32(__s0_613), splat_laneq_s32(__s1_613, __p2_613));
  __ret_613;
})

#define vmull_high_laneq_s16(__p0_615, __p1_615, __p2_615) __extension__ ({
  int32x4_t __ret_615;
  int16x8_t __s0_615 = __p0_615;
  int16x8_t __s1_615 = __p1_615;
  __ret_615 = vmull_s16(vget_high_s16(__s0_615), splat_laneq_s16(__s1_615, __p2_615));
  __ret_615;
})

__ai __attribute__((target("neon")))
uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1)

__ai __attribute__((target("neon")))
uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1)

#define vmull_laneq_u32(__p0_617, __p1_617, __p2_617) __extension__ ({
  uint64x2_t __ret_617;
  uint32x2_t __s0_617 = __p0_617;
  uint32x4_t __s1_617 = __p1_617;
  __ret_617 = vmull_u32(__s0_617, splat_laneq_u32(__s1_617, __p2_617));
  __ret_617;
})

#define vmull_laneq_u16(__p0_619, __p1_619, __p2_619) __extension__ ({
  uint32x4_t __ret_619;
  uint16x4_t __s0_619 = __p0_619;
  uint16x8_t __s1_619 = __p1_619;
  __ret_619 = vmull_u16(__s0_619, splat_laneq_u16(__s1_619, __p2_619));
  __ret_619;
})

#define vmull_laneq_s32(__p0_621, __p1_621, __p2_621) __extension__ ({
  int64x2_t __ret_621;
  int32x2_t __s0_621 = __p0_621;
  int32x4_t __s1_621 = __p1_621;
  __ret_621 = vmull_s32(__s0_621, splat_laneq_s32(__s1_621, __p2_621));
  __ret_621;
})

#define vmull_laneq_s16(__p0_623, __p1_623, __p2_623) __extension__ ({
  int32x4_t __ret_623;
  int16x4_t __s0_623 = __p0_623;
  int16x8_t __s1_623 = __p1_623;
  __ret_623 = vmull_s16(__s0_623, splat_laneq_s16(__s1_623, __p2_623));
  __ret_623;
})

__ai __attribute__((target("neon")))
float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1)

__ai __attribute__((target("neon")))
float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1)

__ai __attribute__((target("neon")))
float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1)

__ai __attribute__((target("neon")))
float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1)

__ai __attribute__((target("neon")))
float64_t vmulxd_f64(float64_t __p0, float64_t __p1)

__ai __attribute__((target("neon")))
float32_t vmulxs_f32(float32_t __p0, float32_t __p1)

#define vmulxd_lane_f64(__p0_625, __p1_625, __p2_625) __extension__ ({
  float64_t __ret_625;
  float64_t __s0_625 = __p0_625;
  float64x1_t __s1_625 = __p1_625;
  __ret_625 = vmulxd_f64(__s0_625, vget_lane_f64(__s1_625, __p2_625));
  __ret_625;
})

#define vmulxs_lane_f32(__p0_626, __p1_626, __p2_626) __extension__ ({
  float32_t __ret_626;
  float32_t __s0_626 = __p0_626;
  float32x2_t __s1_626 = __p1_626;
  __ret_626 = vmulxs_f32(__s0_626, vget_lane_f32(__s1_626, __p2_626));
  __ret_626;
})

#define vmulxq_lane_f64(__p0_628, __p1_628, __p2_628) __extension__ ({
  float64x2_t __ret_628;
  float64x2_t __s0_628 = __p0_628;
  float64x1_t __s1_628 = __p1_628;
  __ret_628 = vmulxq_f64(__s0_628, splatq_lane_f64(__s1_628, __p2_628));
  __ret_628;
})

#define vmulxq_lane_f32(__p0_630, __p1_630, __p2_630) __extension__ ({
  float32x4_t __ret_630;
  float32x4_t __s0_630 = __p0_630;
  float32x2_t __s1_630 = __p1_630;
  __ret_630 = vmulxq_f32(__s0_630, splatq_lane_f32(__s1_630, __p2_630));
  __ret_630;
})

#define vmulx_lane_f32(__p0_632, __p1_632, __p2_632) __extension__ ({
  float32x2_t __ret_632;
  float32x2_t __s0_632 = __p0_632;
  float32x2_t __s1_632 = __p1_632;
  __ret_632 = vmulx_f32(__s0_632, splat_lane_f32(__s1_632, __p2_632));
  __ret_632;
})

#define vmulxd_laneq_f64(__p0_634, __p1_634, __p2_634) __extension__ ({
  float64_t __ret_634;
  float64_t __s0_634 = __p0_634;
  float64x2_t __s1_634 = __p1_634;
  __ret_634 = vmulxd_f64(__s0_634, vgetq_lane_f64(__s1_634, __p2_634));
  __ret_634;
})

#define vmulxs_laneq_f32(__p0_636, __p1_636, __p2_636) __extension__ ({
  float32_t __ret_636;
  float32_t __s0_636 = __p0_636;
  float32x4_t __s1_636 = __p1_636;
  __ret_636 = vmulxs_f32(__s0_636, vgetq_lane_f32(__s1_636, __p2_636));
  __ret_636;
})

#define vmulxq_laneq_f64(__p0_638, __p1_638, __p2_638) __extension__ ({
  float64x2_t __ret_638;
  float64x2_t __s0_638 = __p0_638;
  float64x2_t __s1_638 = __p1_638;
  __ret_638 = vmulxq_f64(__s0_638, splatq_laneq_f64(__s1_638, __p2_638));
  __ret_638;
})

#define vmulxq_laneq_f32(__p0_640, __p1_640, __p2_640) __extension__ ({
  float32x4_t __ret_640;
  float32x4_t __s0_640 = __p0_640;
  float32x4_t __s1_640 = __p1_640;
  __ret_640 = vmulxq_f32(__s0_640, splatq_laneq_f32(__s1_640, __p2_640));
  __ret_640;
})

#define vmulx_laneq_f32(__p0_642, __p1_642, __p2_642) __extension__ ({
  float32x2_t __ret_642;
  float32x2_t __s0_642 = __p0_642;
  float32x4_t __s1_642 = __p1_642;
  __ret_642 = vmulx_f32(__s0_642, splat_laneq_f32(__s1_642, __p2_642));
  __ret_642;
})

__ai __attribute__((target("neon")))
int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2)

#define vqdmlal_high_lane_s32(__p0_644, __p1_644, __p2_644, __p3_644) __extension__ ({
  int64x2_t __ret_644;
  int64x2_t __s0_644 = __p0_644;
  int32x4_t __s1_644 = __p1_644;
  int32x2_t __s2_644 = __p2_644;
  __ret_644 = vqdmlal_s32(__s0_644, vget_high_s32(__s1_644), splat_lane_s32(__s2_644, __p3_644));
  __ret_644;
})

#define vqdmlal_high_lane_s16(__p0_646, __p1_646, __p2_646, __p3_646) __extension__ ({
  int32x4_t __ret_646;
  int32x4_t __s0_646 = __p0_646;
  int16x8_t __s1_646 = __p1_646;
  int16x4_t __s2_646 = __p2_646;
  __ret_646 = vqdmlal_s16(__s0_646, vget_high_s16(__s1_646), splat_lane_s16(__s2_646, __p3_646));
  __ret_646;
})

#define vqdmlal_high_laneq_s32(__p0_648, __p1_648, __p2_648, __p3_648) __extension__ ({
  int64x2_t __ret_648;
  int64x2_t __s0_648 = __p0_648;
  int32x4_t __s1_648 = __p1_648;
  int32x4_t __s2_648 = __p2_648;
  __ret_648 = vqdmlal_s32(__s0_648, vget_high_s32(__s1_648), splat_laneq_s32(__s2_648, __p3_648));
  __ret_648;
})

#define vqdmlal_high_laneq_s16(__p0_650, __p1_650, __p2_650, __p3_650) __extension__ ({
  int32x4_t __ret_650;
  int32x4_t __s0_650 = __p0_650;
  int16x8_t __s1_650 = __p1_650;
  int16x8_t __s2_650 = __p2_650;
  __ret_650 = vqdmlal_s16(__s0_650, vget_high_s16(__s1_650), splat_laneq_s16(__s2_650, __p3_650));
  __ret_650;
})

__ai __attribute__((target("neon")))
int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2)

#define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({
  int64_t __ret;
  int64_t __s0 = __p0;
  int32_t __s1 = __p1;
  int32x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(int64_t, __builtin_neon_vqdmlals_lane_s32(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({
  int32_t __ret;
  int32_t __s0 = __p0;
  int16_t __s1 = __p1;
  int16x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(int32_t, __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({
  int64_t __ret;
  int64_t __s0 = __p0;
  int32_t __s1 = __p1;
  int32x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(int64_t, __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({
  int32_t __ret;
  int32_t __s0 = __p0;
  int16_t __s1 = __p1;
  int16x8_t __s2 = __p2;
  __ret = __builtin_bit_cast(int32_t, __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlal_laneq_s32(__p0_652, __p1_652, __p2_652, __p3_652) __extension__ ({
  int64x2_t __ret_652;
  int64x2_t __s0_652 = __p0_652;
  int32x2_t __s1_652 = __p1_652;
  int32x4_t __s2_652 = __p2_652;
  __ret_652 = vqdmlal_s32(__s0_652, __s1_652, splat_laneq_s32(__s2_652, __p3_652));
  __ret_652;
})

#define vqdmlal_laneq_s16(__p0_654, __p1_654, __p2_654, __p3_654) __extension__ ({
  int32x4_t __ret_654;
  int32x4_t __s0_654 = __p0_654;
  int16x4_t __s1_654 = __p1_654;
  int16x8_t __s2_654 = __p2_654;
  __ret_654 = vqdmlal_s16(__s0_654, __s1_654, splat_laneq_s16(__s2_654, __p3_654));
  __ret_654;
})

__ai __attribute__((target("neon")))
int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2)

#define vqdmlsl_high_lane_s32(__p0_656, __p1_656, __p2_656, __p3_656) __extension__ ({
  int64x2_t __ret_656;
  int64x2_t __s0_656 = __p0_656;
  int32x4_t __s1_656 = __p1_656;
  int32x2_t __s2_656 = __p2_656;
  __ret_656 = vqdmlsl_s32(__s0_656, vget_high_s32(__s1_656), splat_lane_s32(__s2_656, __p3_656));
  __ret_656;
})

#define vqdmlsl_high_lane_s16(__p0_658, __p1_658, __p2_658, __p3_658) __extension__ ({
  int32x4_t __ret_658;
  int32x4_t __s0_658 = __p0_658;
  int16x8_t __s1_658 = __p1_658;
  int16x4_t __s2_658 = __p2_658;
  __ret_658 = vqdmlsl_s16(__s0_658, vget_high_s16(__s1_658), splat_lane_s16(__s2_658, __p3_658));
  __ret_658;
})

#define vqdmlsl_high_laneq_s32(__p0_660, __p1_660, __p2_660, __p3_660) __extension__ ({
  int64x2_t __ret_660;
  int64x2_t __s0_660 = __p0_660;
  int32x4_t __s1_660 = __p1_660;
  int32x4_t __s2_660 = __p2_660;
  __ret_660 = vqdmlsl_s32(__s0_660, vget_high_s32(__s1_660), splat_laneq_s32(__s2_660, __p3_660));
  __ret_660;
})

#define vqdmlsl_high_laneq_s16(__p0_662, __p1_662, __p2_662, __p3_662) __extension__ ({
  int32x4_t __ret_662;
  int32x4_t __s0_662 = __p0_662;
  int16x8_t __s1_662 = __p1_662;
  int16x8_t __s2_662 = __p2_662;
  __ret_662 = vqdmlsl_s16(__s0_662, vget_high_s16(__s1_662), splat_laneq_s16(__s2_662, __p3_662));
  __ret_662;
})

__ai __attribute__((target("neon")))
int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2)

#define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({
  int64_t __ret;
  int64_t __s0 = __p0;
  int32_t __s1 = __p1;
  int32x2_t __s2 = __p2;
  __ret = __builtin_bit_cast(int64_t, __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({
  int32_t __ret;
  int32_t __s0 = __p0;
  int16_t __s1 = __p1;
  int16x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(int32_t, __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({
  int64_t __ret;
  int64_t __s0 = __p0;
  int32_t __s1 = __p1;
  int32x4_t __s2 = __p2;
  __ret = __builtin_bit_cast(int64_t, __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({
  int32_t __ret;
  int32_t __s0 = __p0;
  int16_t __s1 = __p1;
  int16x8_t __s2 = __p2;
  __ret = __builtin_bit_cast(int32_t, __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, __s2, __p3));
  __ret;
})

#define vqdmlsl_laneq_s32(__p0_664, __p1_664, __p2_664, __p3_664) __extension__ ({
  int64x2_t __ret_664;
  int64x2_t __s0_664 = __p0_664;
  int32x2_t __s1_664 = __p1_664;
  int32x4_t __s2_664 = __p2_664;
  __ret_664 = vqdmlsl_s32(__s0_664, __s1_664, splat_laneq_s32(__s2_664, __p3_664));
  __ret_664;
})

#define vqdmlsl_laneq_s16(__p0_666, __p1_666, __p2_666, __p3_666) __extension__ ({
  int32x4_t __ret_666;
  int32x4_t __s0_666 = __p0_666;
  int16x4_t __s1_666 = __p1_666;
  int16x8_t __s2_666 = __p2_666;
  __ret_666 = vqdmlsl_s16(__s0_666, __s1_666, splat_laneq_s16(__s2_666, __p3_666));
  __ret_666;
})

__ai __attribute__((target("neon")))
int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1)

#define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({
  int32x4_t __ret;
  int32x4_t __s0 = __p0;
  int32x2_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x4_t, __builtin_neon_vqdmulhq_lane_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 2));
  __ret;
})

#define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({
  int16x8_t __ret;
  int16x8_t __s0 = __p0;
  int16x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x8_t, __builtin_neon_vqdmulhq_lane_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 1));
  __ret;
})

#define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({
  int32x2_t __ret;
  int32x2_t __s0 = __p0;
  int32x2_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x2_t, __builtin_neon_vqdmulh_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 2));
  __ret;
})

#define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({
  int16x4_t __ret;
  int16x4_t __s0 = __p0;
  int16x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x4_t, __builtin_neon_vqdmulh_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 1));
  __ret;
})

#define vqdmulhs_lane_s32(__p0_668, __p1_668, __p2_668) __extension__ ({
  int32_t __ret_668;
  int32_t __s0_668 = __p0_668;
  int32x2_t __s1_668 = __p1_668;
  __ret_668 = vqdmulhs_s32(__s0_668, vget_lane_s32(__s1_668, __p2_668));
  __ret_668;
})

#define vqdmulhh_lane_s16(__p0_670, __p1_670, __p2_670) __extension__ ({
  int16_t __ret_670;
  int16_t __s0_670 = __p0_670;
  int16x4_t __s1_670 = __p1_670;
  __ret_670 = vqdmulhh_s16(__s0_670, vget_lane_s16(__s1_670, __p2_670));
  __ret_670;
})

#define vqdmulhs_laneq_s32(__p0_672, __p1_672, __p2_672) __extension__ ({
  int32_t __ret_672;
  int32_t __s0_672 = __p0_672;
  int32x4_t __s1_672 = __p1_672;
  __ret_672 = vqdmulhs_s32(__s0_672, vgetq_lane_s32(__s1_672, __p2_672));
  __ret_672;
})

#define vqdmulhh_laneq_s16(__p0_674, __p1_674, __p2_674) __extension__ ({
  int16_t __ret_674;
  int16_t __s0_674 = __p0_674;
  int16x8_t __s1_674 = __p1_674;
  __ret_674 = vqdmulhh_s16(__s0_674, vgetq_lane_s16(__s1_674, __p2_674));
  __ret_674;
})

#define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({
  int32x4_t __ret;
  int32x4_t __s0 = __p0;
  int32x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x4_t, __builtin_neon_vqdmulhq_laneq_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 34));
  __ret;
})

#define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({
  int16x8_t __ret;
  int16x8_t __s0 = __p0;
  int16x8_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x8_t, __builtin_neon_vqdmulhq_laneq_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 33));
  __ret;
})

#define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({
  int32x2_t __ret;
  int32x2_t __s0 = __p0;
  int32x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x2_t, __builtin_neon_vqdmulh_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 2));
  __ret;
})

#define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({
  int16x4_t __ret;
  int16x4_t __s0 = __p0;
  int16x8_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x4_t, __builtin_neon_vqdmulh_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 1));
  __ret;
})

__ai __attribute__((target("neon")))
int64_t vqdmulls_s32(int32_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int32_t vqdmullh_s16(int16_t __p0, int16_t __p1)

__ai __attribute__((target("neon")))
int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1)

#define vqdmull_high_lane_s32(__p0_676, __p1_676, __p2_676) __extension__ ({
  int64x2_t __ret_676;
  int32x4_t __s0_676 = __p0_676;
  int32x2_t __s1_676 = __p1_676;
  __ret_676 = vqdmull_s32(vget_high_s32(__s0_676), splat_lane_s32(__s1_676, __p2_676));
  __ret_676;
})

#define vqdmull_high_lane_s16(__p0_678, __p1_678, __p2_678) __extension__ ({
  int32x4_t __ret_678;
  int16x8_t __s0_678 = __p0_678;
  int16x4_t __s1_678 = __p1_678;
  __ret_678 = vqdmull_s16(vget_high_s16(__s0_678), splat_lane_s16(__s1_678, __p2_678));
  __ret_678;
})

#define vqdmull_high_laneq_s32(__p0_680, __p1_680, __p2_680) __extension__ ({
  int64x2_t __ret_680;
  int32x4_t __s0_680 = __p0_680;
  int32x4_t __s1_680 = __p1_680;
  __ret_680 = vqdmull_s32(vget_high_s32(__s0_680), splat_laneq_s32(__s1_680, __p2_680));
  __ret_680;
})

#define vqdmull_high_laneq_s16(__p0_682, __p1_682, __p2_682) __extension__ ({
  int32x4_t __ret_682;
  int16x8_t __s0_682 = __p0_682;
  int16x8_t __s1_682 = __p1_682;
  __ret_682 = vqdmull_s16(vget_high_s16(__s0_682), splat_laneq_s16(__s1_682, __p2_682));
  __ret_682;
})

__ai __attribute__((target("neon")))
int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1)

#define vqdmulls_lane_s32(__p0_684, __p1_684, __p2_684) __extension__ ({
  int64_t __ret_684;
  int32_t __s0_684 = __p0_684;
  int32x2_t __s1_684 = __p1_684;
  __ret_684 = vqdmulls_s32(__s0_684, vget_lane_s32(__s1_684, __p2_684));
  __ret_684;
})

#define vqdmullh_lane_s16(__p0_686, __p1_686, __p2_686) __extension__ ({
  int32_t __ret_686;
  int16_t __s0_686 = __p0_686;
  int16x4_t __s1_686 = __p1_686;
  __ret_686 = vqdmullh_s16(__s0_686, vget_lane_s16(__s1_686, __p2_686));
  __ret_686;
})

#define vqdmulls_laneq_s32(__p0_688, __p1_688, __p2_688) __extension__ ({
  int64_t __ret_688;
  int32_t __s0_688 = __p0_688;
  int32x4_t __s1_688 = __p1_688;
  __ret_688 = vqdmulls_s32(__s0_688, vgetq_lane_s32(__s1_688, __p2_688));
  __ret_688;
})

#define vqdmullh_laneq_s16(__p0_690, __p1_690, __p2_690) __extension__ ({
  int32_t __ret_690;
  int16_t __s0_690 = __p0_690;
  int16x8_t __s1_690 = __p1_690;
  __ret_690 = vqdmullh_s16(__s0_690, vgetq_lane_s16(__s1_690, __p2_690));
  __ret_690;
})

#define vqdmull_laneq_s32(__p0_692, __p1_692, __p2_692) __extension__ ({
  int64x2_t __ret_692;
  int32x2_t __s0_692 = __p0_692;
  int32x4_t __s1_692 = __p1_692;
  __ret_692 = vqdmull_s32(__s0_692, splat_laneq_s32(__s1_692, __p2_692));
  __ret_692;
})

#define vqdmull_laneq_s16(__p0_694, __p1_694, __p2_694) __extension__ ({
  int32x4_t __ret_694;
  int16x4_t __s0_694 = __p0_694;
  int16x8_t __s1_694 = __p1_694;
  __ret_694 = vqdmull_s16(__s0_694, splat_laneq_s16(__s1_694, __p2_694));
  __ret_694;
})

__ai __attribute__((target("neon")))
int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1)

__ai __attribute__((target("neon")))
int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1)

#define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({
  int32x4_t __ret;
  int32x4_t __s0 = __p0;
  int32x2_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x4_t, __builtin_neon_vqrdmulhq_lane_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 2));
  __ret;
})

#define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({
  int16x8_t __ret;
  int16x8_t __s0 = __p0;
  int16x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x8_t, __builtin_neon_vqrdmulhq_lane_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 1));
  __ret;
})

#define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({
  int32x2_t __ret;
  int32x2_t __s0 = __p0;
  int32x2_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x2_t, __builtin_neon_vqrdmulh_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 2));
  __ret;
})

#define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({
  int16x4_t __ret;
  int16x4_t __s0 = __p0;
  int16x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x4_t, __builtin_neon_vqrdmulh_lane_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x8_t, __s1), __p2, 1));
  __ret;
})

#define vqrdmulhs_lane_s32(__p0_696, __p1_696, __p2_696) __extension__ ({
  int32_t __ret_696;
  int32_t __s0_696 = __p0_696;
  int32x2_t __s1_696 = __p1_696;
  __ret_696 = vqrdmulhs_s32(__s0_696, vget_lane_s32(__s1_696, __p2_696));
  __ret_696;
})

#define vqrdmulhh_lane_s16(__p0_698, __p1_698, __p2_698) __extension__ ({
  int16_t __ret_698;
  int16_t __s0_698 = __p0_698;
  int16x4_t __s1_698 = __p1_698;
  __ret_698 = vqrdmulhh_s16(__s0_698, vget_lane_s16(__s1_698, __p2_698));
  __ret_698;
})

#define vqrdmulhs_laneq_s32(__p0_700, __p1_700, __p2_700) __extension__ ({
  int32_t __ret_700;
  int32_t __s0_700 = __p0_700;
  int32x4_t __s1_700 = __p1_700;
  __ret_700 = vqrdmulhs_s32(__s0_700, vgetq_lane_s32(__s1_700, __p2_700));
  __ret_700;
})

#define vqrdmulhh_laneq_s16(__p0_702, __p1_702, __p2_702) __extension__ ({
  int16_t __ret_702;
  int16_t __s0_702 = __p0_702;
  int16x8_t __s1_702 = __p1_702;
  __ret_702 = vqrdmulhh_s16(__s0_702, vgetq_lane_s16(__s1_702, __p2_702));
  __ret_702;
})

#define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({
  int32x4_t __ret;
  int32x4_t __s0 = __p0;
  int32x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x4_t, __builtin_neon_vqrdmulhq_laneq_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 34));
  __ret;
})

#define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({
  int16x8_t __ret;
  int16x8_t __s0 = __p0;
  int16x8_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x8_t, __builtin_neon_vqrdmulhq_laneq_v(__builtin_bit_cast(int8x16_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 33));
  __ret;
})

#define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({
  int32x2_t __ret;
  int32x2_t __s0 = __p0;
  int32x4_t __s1 = __p1;
  __ret = __builtin_bit_cast(int32x2_t, __builtin_neon_vqrdmulh_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 2));
  __ret;
})

#define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({
  int16x4_t __ret;
  int16x4_t __s0 = __p0;
  int16x8_t __s1 = __p1;
  __ret = __builtin_bit_cast(int16x4_t, __builtin_neon_vqrdmulh_laneq_v(__builtin_bit_cast(int8x8_t, __s0), __builtin_bit_cast(int8x16_t, __s1), __p2, 1));
  __ret;
})

__ai __attribute__((target("v8.1a,neon")))
int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2)

#define vqrdmlahs_lane_s32(__p0_776, __p1_776, __p2_776, __p3_776) __extension__ ({
  int32_t __ret_776;
  int32_t __s0_776 = __p0_776;
  int32_t __s1_776 = __p1_776;
  int32x2_t __s2_776 = __p2_776;
  __ret_776 = vqrdmlahs_s32(__s0_776, __s1_776, vget_lane_s32(__s2_776, __p3_776));
  __ret_776;
})

#define vqrdmlahh_lane_s16(__p0_778, __p1_778, __p2_778, __p3_778) __extension__ ({
  int16_t __ret_778;
  int16_t __s0_778 = __p0_778;
  int16_t __s1_778 = __p1_778;
  int16x4_t __s2_778 = __p2_778;
  __ret_778 = vqrdmlahh_s16(__s0_778, __s1_778, vget_lane_s16(__s2_778, __p3_778));
  __ret_778;
})

#define vqrdmlahs_laneq_s32(__p0_780, __p1_780, __p2_780, __p3_780) __extension__ ({
  int32_t __ret_780;
  int32_t __s0_780 = __p0_780;
  int32_t __s1_780 = __p1_780;
  int32x4_t __s2_780 = __p2_780;
  __ret_780 = vqrdmlahs_s32(__s0_780, __s1_780, vgetq_lane_s32(__s2_780, __p3_780));
  __ret_780;
})

#define vqrdmlahh_laneq_s16(__p0_782, __p1_782, __p2_782, __p3_782) __extension__ ({
  int16_t __ret_782;
  int16_t __s0_782 = __p0_782;
  int16_t __s1_782 = __p1_782;
  int16x8_t __s2_782 = __p2_782;
  __ret_782 = vqrdmlahh_s16(__s0_782, __s1_782, vgetq_lane_s16(__s2_782, __p3_782));
  __ret_782;
})

#define vqrdmlahq_laneq_s32(__p0_784, __p1_784, __p2_784, __p3_784) __extension__ ({
  int32x4_t __ret_784;
  int32x4_t __s0_784 = __p0_784;
  int32x4_t __s1_784 = __p1_784;
  int32x4_t __s2_784 = __p2_784;
  __ret_784 = vqrdmlahq_s32(__s0_784, __s1_784, splatq_laneq_s32(__s2_784, __p3_784));
  __ret_784;
})

#define vqrdmlahq_laneq_s16(__p0_786, __p1_786, __p2_786, __p3_786) __extension__ ({
  int16x8_t __ret_786;
  int16x8_t __s0_786 = __p0_786;
  int16x8_t __s1_786 = __p1_786;
  int16x8_t __s2_786 = __p2_786;
  __ret_786 = vqrdmlahq_s16(__s0_786, __s1_786, splatq_laneq_s16(__s2_786, __p3_786));
  __ret_786;
})

#define vqrdmlah_laneq_s32(__p0_788, __p1_788, __p2_788, __p3_788) __extension__ ({
  int32x2_t __ret_788;
  int32x2_t __s0_788 = __p0_788;
  int32x2_t __s1_788 = __p1_788;
  int32x4_t __s2_788 = __p2_788;
  __ret_788 = vqrdmlah_s32(__s0_788, __s1_788, splat_laneq_s32(__s2_788, __p3_788));
  __ret_788;
})

#define vqrdmlah_laneq_s16(__p0_790, __p1_790, __p2_790, __p3_790) __extension__ ({
  int16x4_t __ret_790;
  int16x4_t __s0_790 = __p0_790;
  int16x4_t __s1_790 = __p1_790;
  int16x8_t __s2_790 = __p2_790;
  __ret_790 = vqrdmlah_s16(__s0_790, __s1_790, splat_laneq_s16(__s2_790, __p3_790));
  __ret_790;
})

__ai __attribute__((target("v8.1a,neon")))
int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2)

__ai __attribute__((target("v8.1a,neon")))
int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2)

#define vqrdmlshs_lane_s32(__p0_792, __p1_792, __p2_792, __p3_792) __extension__ ({
  int32_t __ret_792;
  int32_t __s0_792 = __p0_792;
  int32_t __s1_792 = __p1_792;
  int32x2_t __s2_792 = __p2_792;
  __ret_792 = vqrdmlshs_s32(__s0_792, __s1_792, vget_lane_s32(__s2_792, __p3_792));
  __ret_792;
})

#define vqrdmlshh_lane_s16(__p0_794, __p1_794, __p2_794, __p3_794) __extension__ ({
  int16_t __ret_794;
  int16_t __s0_794 = __p0_794;
  int16_t __s1_794 = __p1_794;
  int16x4_t __s2_794 = __p2_794;
  __ret_794 = vqrdmlshh_s16(__s0_794, __s1_794, vget_lane_s16(__s2_794, __p3_794));
  __ret_794;
})

#define vqrdmlshs_laneq_s32(__p0_796, __p1_796, __p2_796, __p3_796) __extension__ ({
  int32_t __ret_796;
  int32_t __s0_796 = __p0_796;
  int32_t __s1_796 = __p1_796;
  int32x4_t __s2_796 = __p2_796;
  __ret_796 = vqrdmlshs_s32(__s0_796, __s1_796, vgetq_lane_s32(__s2_796, __p3_796));
  __ret_796;
})

#define vqrdmlshh_laneq_s16(__p0_798, __p1_798, __p2_798, __p3_798) __extension__ ({
  int16_t __ret_798;
  int16_t __s0_798 = __p0_798;
  int16_t __s1_798 = __p1_798;
  int16x8_t __s2_798 = __p2_798;
  __ret_798 = vqrdmlshh_s16(__s0_798, __s1_798, vgetq_lane_s16(__s2_798, __p3_798));
  __ret_798;
})

#define vqrdmlshq_laneq_s32(__p0_800, __p1_800, __p2_800, __p3_800) __extension__ ({
  int32x4_t __ret_800;
  int32x4_t __s0_800 = __p0_800;
  int32x4_t __s1_800 = __p1_800;
  int32x4_t __s2_800 = __p2_800;
  __ret_800 = vqrdmlshq_s32(__s0_800, __s1_800, splatq_laneq_s32(__s2_800, __p3_800));
  __ret_800;
})

#define vqrdmlshq_laneq_s16(__p0_802, __p1_802, __p2_802, __p3_802) __extension__ ({
  int16x8_t __ret_802;
  int16x8_t __s0_802 = __p0_802;
  int16x8_t __s1_802 = __p1_802;
  int16x8_t __s2_802 = __p2_802;
  __ret_802 = vqrdmlshq_s16(__s0_802, __s1_802, splatq_laneq_s16(__s2_802, __p3_802));
  __ret_802;
})

#define vqrdmlsh_laneq_s32(__p0_804, __p1_804, __p2_804, __p3_804) __extension__ ({
  int32x2_t __ret_804;
  int32x2_t __s0_804 = __p0_804;
  int32x2_t __s1_804 = __p1_804;
  int32x4_t __s2_804 = __p2_804;
  __ret_804 = vqrdmlsh_s32(__s0_804, __s1_804, splat_laneq_s32(__s2_804, __p3_804));
  __ret_804;
})

#define vqrdmlsh_laneq_s16(__p0_806, __p1_806, __p2_806, __p3_806) __extension__ ({
  int16x4_t __ret_806;
  int16x4_t __s0_806 = __p0_806;
  int16x4_t __s1_806 = __p1_806;
  int16x8_t __s2_806 = __p2_806;
  __ret_806 = vqrdmlsh_s16(__s0_806, __s1_806, splat_laneq_s16(__s2_806, __p3_806));
  __ret_806;
})

#define vbfmlalbq_lane_f32(__p0_816, __p1_816, __p2_816, __p3_816) __extension__ ({
  float32x4_t __ret_816;
  float32x4_t __s0_816 = __p0_816;
  bfloat16x8_t __s1_816 = __p1_816;
  bfloat16x4_t __s2_816 = __p2_816;
  __ret_816 = vbfmlalbq_f32(__s0_816, __s1_816, (bfloat16x8_t) {vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816), vget_lane_bf16(__s2_816, __p3_816)});
  __ret_816;
})

#define vbfmlalbq_laneq_f32(__p0_818, __p1_818, __p2_818, __p3_818) __extension__ ({
  float32x4_t __ret_818;
  float32x4_t __s0_818 = __p0_818;
  bfloat16x8_t __s1_818 = __p1_818;
  bfloat16x8_t __s2_818 = __p2_818;
  __ret_818 = vbfmlalbq_f32(__s0_818, __s1_818, (bfloat16x8_t) {vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818), vgetq_lane_bf16(__s2_818, __p3_818)});
  __ret_818;
})

#define vbfmlaltq_lane_f32(__p0_820, __p1_820, __p2_820, __p3_820) __extension__ ({
  float32x4_t __ret_820;
  float32x4_t __s0_820 = __p0_820;
  bfloat16x8_t __s1_820 = __p1_820;
  bfloat16x4_t __s2_820 = __p2_820;
  __ret_820 = vbfmlaltq_f32(__s0_820, __s1_820, (bfloat16x8_t) {vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820), vget_lane_bf16(__s2_820, __p3_820)});
  __ret_820;
})

#define vbfmlaltq_laneq_f32(__p0_822, __p1_822, __p2_822, __p3_822) __extension__ ({
  float32x4_t __ret_822;
  float32x4_t __s0_822 = __p0_822;
  bfloat16x8_t __s1_822 = __p1_822;
  bfloat16x8_t __s2_822 = __p2_822;
  __ret_822 = vbfmlaltq_f32(__s0_822, __s1_822, (bfloat16x8_t) {vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822), vgetq_lane_bf16(__s2_822, __p3_822)});
  __ret_822;
})

#define vmulq_lane_f16(__p0_835, __p1_835, __p2_835) __extension__ ({
  float16x8_t __ret_835;
  float16x8_t __s0_835 = __p0_835;
  float16x4_t __s1_835 = __p1_835;
  __ret_835 = __s0_835 * splatq_lane_f16(__s1_835, __p2_835);
  __ret_835;
})

#define vmul_lane_f16(__p0_837, __p1_837, __p2_837) __extension__ ({
  float16x4_t __ret_837;
  float16x4_t __s0_837 = __p0_837;
  float16x4_t __s1_837 = __p1_837;
  __ret_837 = __s0_837 * splat_lane_f16(__s1_837, __p2_837);
  __ret_837;
})

__ai __attribute__((target("neon")))
uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2)

__ai __attribute__((target("neon")))
uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vmlal_lane_u32(__p0_853, __p1_853, __p2_853, __p3_853) __extension__ ({
  uint64x2_t __ret_853;
  uint64x2_t __s0_853 = __p0_853;
  uint32x2_t __s1_853 = __p1_853;
  uint32x2_t __s2_853 = __p2_853;
  __ret_853 = __s0_853 + vmull_u32(__s1_853, splat_lane_u32(__s2_853, __p3_853));
  __ret_853;
})

#define vmlal_lane_u16(__p0_855, __p1_855, __p2_855, __p3_855) __extension__ ({
  uint32x4_t __ret_855;
  uint32x4_t __s0_855 = __p0_855;
  uint16x4_t __s1_855 = __p1_855;
  uint16x4_t __s2_855 = __p2_855;
  __ret_855 = __s0_855 + vmull_u16(__s1_855, splat_lane_u16(__s2_855, __p3_855));
  __ret_855;
})

#define vmlal_lane_s32(__p0_857, __p1_857, __p2_857, __p3_857) __extension__ ({
  int64x2_t __ret_857;
  int64x2_t __s0_857 = __p0_857;
  int32x2_t __s1_857 = __p1_857;
  int32x2_t __s2_857 = __p2_857;
  __ret_857 = __s0_857 + vmull_s32(__s1_857, splat_lane_s32(__s2_857, __p3_857));
  __ret_857;
})

#define vmlal_lane_s16(__p0_859, __p1_859, __p2_859, __p3_859) __extension__ ({
  int32x4_t __ret_859;
  int32x4_t __s0_859 = __p0_859;
  int16x4_t __s1_859 = __p1_859;
  int16x4_t __s2_859 = __p2_859;
  __ret_859 = __s0_859 + vmull_s16(__s1_859, splat_lane_s16(__s2_859, __p3_859));
  __ret_859;
})

__ai __attribute__((target("neon")))
uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2)

__ai __attribute__((target("neon")))
uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2)

#define vmlsl_lane_u32(__p0_861, __p1_861, __p2_861, __p3_861) __extension__ ({
  uint64x2_t __ret_861;
  uint64x2_t __s0_861 = __p0_861;
  uint32x2_t __s1_861 = __p1_861;
  uint32x2_t __s2_861 = __p2_861;
  __ret_861 = __s0_861 - vmull_u32(__s1_861, splat_lane_u32(__s2_861, __p3_861));
  __ret_861;
})

#define vmlsl_lane_u16(__p0_863, __p1_863, __p2_863, __p3_863) __extension__ ({
  uint32x4_t __ret_863;
  uint32x4_t __s0_863 = __p0_863;
  uint16x4_t __s1_863 = __p1_863;
  uint16x4_t __s2_863 = __p2_863;
  __ret_863 = __s0_863 - vmull_u16(__s1_863, splat_lane_u16(__s2_863, __p3_863));
  __ret_863;
})

#define vmlsl_lane_s32(__p0_865, __p1_865, __p2_865, __p3_865) __extension__ ({
  int64x2_t __ret_865;
  int64x2_t __s0_865 = __p0_865;
  int32x2_t __s1_865 = __p1_865;
  int32x2_t __s2_865 = __p2_865;
  __ret_865 = __s0_865 - vmull_s32(__s1_865, splat_lane_s32(__s2_865, __p3_865));
  __ret_865;
})

#define vmlsl_lane_s16(__p0_867, __p1_867, __p2_867, __p3_867) __extension__ ({
  int32x4_t __ret_867;
  int32x4_t __s0_867 = __p0_867;
  int16x4_t __s1_867 = __p1_867;
  int16x4_t __s2_867 = __p2_867;
  __ret_867 = __s0_867 - vmull_s16(__s1_867, splat_lane_s16(__s2_867, __p3_867));
  __ret_867;
})

__ai __attribute__((target("neon")))
uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2)

__ai __attribute__((target("aes,neon")))
poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1)

#define vfmlalq_lane_high_f16(__p0_873, __p1_873, __p2_873, __p3_873) __extension__ ({
  float32x4_t __ret_873;
  float32x4_t __s0_873 = __p0_873;
  float16x8_t __s1_873 = __p1_873;
  float16x4_t __s2_873 = __p2_873;
  __ret_873 = vfmlalq_high_f16(__s0_873, __s1_873, (float16x8_t) {vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873)});
  __ret_873;
})

#define vfmlal_lane_high_f16(__p0_875, __p1_875, __p2_875, __p3_875) __extension__ ({
  float32x2_t __ret_875;
  float32x2_t __s0_875 = __p0_875;
  float16x4_t __s1_875 = __p1_875;
  float16x4_t __s2_875 = __p2_875;
  __ret_875 = vfmlal_high_f16(__s0_875, __s1_875, (float16x4_t) {vget_lane_f16(__s2_875, __p3_875), vget_lane_f16(__s2_875, __p3_875), vget_lane_f16(__s2_875, __p3_875), vget_lane_f16(__s2_875, __p3_875)});
  __ret_875;
})

#define vfmlalq_lane_low_f16(__p0_877, __p1_877, __p2_877, __p3_877) __extension__ ({
  float32x4_t __ret_877;
  float32x4_t __s0_877 = __p0_877;
  float16x8_t __s1_877 = __p1_877;
  float16x4_t __s2_877 = __p2_877;
  __ret_877 = vfmlalq_low_f16(__s0_877, __s1_877, (float16x8_t) {vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877), vget_lane_f16(__s2_877, __p3_877)});
  __ret_877;
})

#define vfmlal_lane_low_f16(__p0_879, __p1_879, __p2_879, __p3_879) __extension__ ({
  float32x2_t __ret_879;
  float32x2_t __s0_879 = __p0_879;
  float16x4_t __s1_879 = __p1_879;
  float16x4_t __s2_879 = __p2_879;
  __ret_879 = vfmlal_low_f16(__s0_879, __s1_879, (float16x4_t) {vget_lane_f16(__s2_879, __p3_879), vget_lane_f16(__s2_879, __p3_879), vget_lane_f16(__s2_879, __p3_879), vget_lane_f16(__s2_879, __p3_879)});
  __ret_879;
})

#define vfmlalq_laneq_high_f16(__p0_881, __p1_881, __p2_881, __p3_881) __extension__ ({
  float32x4_t __ret_881;
  float32x4_t __s0_881 = __p0_881;
  float16x8_t __s1_881 = __p1_881;
  float16x8_t __s2_881 = __p2_881;
  __ret_881 = vfmlalq_high_f16(__s0_881, __s1_881, (float16x8_t) {vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881)});
  __ret_881;
})

#define vfmlal_laneq_high_f16(__p0_883, __p1_883, __p2_883, __p3_883) __extension__ ({
  float32x2_t __ret_883;
  float32x2_t __s0_883 = __p0_883;
  float16x4_t __s1_883 = __p1_883;
  float16x8_t __s2_883 = __p2_883;
  __ret_883 = vfmlal_high_f16(__s0_883, __s1_883, (float16x4_t) {vgetq_lane_f16(__s2_883, __p3_883), vgetq_lane_f16(__s2_883, __p3_883), vgetq_lane_f16(__s2_883, __p3_883), vgetq_lane_f16(__s2_883, __p3_883)});
  __ret_883;
})

#define vfmlalq_laneq_low_f16(__p0_885, __p1_885, __p2_885, __p3_885) __extension__ ({
  float32x4_t __ret_885;
  float32x4_t __s0_885 = __p0_885;
  float16x8_t __s1_885 = __p1_885;
  float16x8_t __s2_885 = __p2_885;
  __ret_885 = vfmlalq_low_f16(__s0_885, __s1_885, (float16x8_t) {vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885), vgetq_lane_f16(__s2_885, __p3_885)});
  __ret_885;
})

#define vfmlal_laneq_low_f16(__p0_887, __p1_887, __p2_887, __p3_887) __extension__ ({
  float32x2_t __ret_887;
  float32x2_t __s0_887 = __p0_887;
  float16x4_t __s1_887 = __p1_887;
  float16x8_t __s2_887 = __p2_887;
  __ret_887 = vfmlal_low_f16(__s0_887, __s1_887, (float16x4_t) {vgetq_lane_f16(__s2_887, __p3_887), vgetq_lane_f16(__s2_887, __p3_887), vgetq_lane_f16(__s2_887, __p3_887), vgetq_lane_f16(__s2_887, __p3_887)});
  __ret_887;
})

#define vfmlslq_lane_high_f16(__p0_889, __p1_889, __p2_889, __p3_889) __extension__ ({
  float32x4_t __ret_889;
  float32x4_t __s0_889 = __p0_889;
  float16x8_t __s1_889 = __p1_889;
  float16x4_t __s2_889 = __p2_889;
  __ret_889 = vfmlslq_high_f16(__s0_889, __s1_889, (float16x8_t) {vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889)});
  __ret_889;
})

#define vfmlsl_lane_high_f16(__p0_891, __p1_891, __p2_891, __p3_891) __extension__ ({
  float32x2_t __ret_891;
  float32x2_t __s0_891 = __p0_891;
  float16x4_t __s1_891 = __p1_891;
  float16x4_t __s2_891 = __p2_891;
  __ret_891 = vfmlsl_high_f16(__s0_891, __s1_891, (float16x4_t) {vget_lane_f16(__s2_891, __p3_891), vget_lane_f16(__s2_891, __p3_891), vget_lane_f16(__s2_891, __p3_891), vget_lane_f16(__s2_891, __p3_891)});
  __ret_891;
})

#define vfmlslq_lane_low_f16(__p0_893, __p1_893, __p2_893, __p3_893) __extension__ ({
  float32x4_t __ret_893;
  float32x4_t __s0_893 = __p0_893;
  float16x8_t __s1_893 = __p1_893;
  float16x4_t __s2_893 = __p2_893;
  __ret_893 = vfmlslq_low_f16(__s0_893, __s1_893, (float16x8_t) {vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893), vget_lane_f16(__s2_893, __p3_893)});
  __ret_893;
})

#define vfmlsl_lane_low_f16(__p0_895, __p1_895, __p2_895, __p3_895) __extension__ ({
  float32x2_t __ret_895;
  float32x2_t __s0_895 = __p0_895;
  float16x4_t __s1_895 = __p1_895;
  float16x4_t __s2_895 = __p2_895;
  __ret_895 = vfmlsl_low_f16(__s0_895, __s1_895, (float16x4_t) {vget_lane_f16(__s2_895, __p3_895), vget_lane_f16(__s2_895, __p3_895), vget_lane_f16(__s2_895, __p3_895), vget_lane_f16(__s2_895, __p3_895)});
  __ret_895;
})

#define vfmlslq_laneq_high_f16(__p0_897, __p1_897, __p2_897, __p3_897) __extension__ ({
  float32x4_t __ret_897;
  float32x4_t __s0_897 = __p0_897;
  float16x8_t __s1_897 = __p1_897;
  float16x8_t __s2_897 = __p2_897;
  __ret_897 = vfmlslq_high_f16(__s0_897, __s1_897, (float16x8_t) {vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897)});
  __ret_897;
})

#define vfmlsl_laneq_high_f16(__p0_899, __p1_899, __p2_899, __p3_899) __extension__ ({
  float32x2_t __ret_899;
  float32x2_t __s0_899 = __p0_899;
  float16x4_t __s1_899 = __p1_899;
  float16x8_t __s2_899 = __p2_899;
  __ret_899 = vfmlsl_high_f16(__s0_899, __s1_899, (float16x4_t) {vgetq_lane_f16(__s2_899, __p3_899), vgetq_lane_f16(__s2_899, __p3_899), vgetq_lane_f16(__s2_899, __p3_899), vgetq_lane_f16(__s2_899, __p3_899)});
  __ret_899;
})

#define vfmlslq_laneq_low_f16(__p0_901, __p1_901, __p2_901, __p3_901) __extension__ ({
  float32x4_t __ret_901;
  float32x4_t __s0_901 = __p0_901;
  float16x8_t __s1_901 = __p1_901;
  float16x8_t __s2_901 = __p2_901;
  __ret_901 = vfmlslq_low_f16(__s0_901, __s1_901, (float16x8_t) {vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901), vgetq_lane_f16(__s2_901, __p3_901)});
  __ret_901;
})

#define vfmlsl_laneq_low_f16(__p0_903, __p1_903, __p2_903, __p3_903) __extension__ ({
  float32x2_t __ret_903;
  float32x2_t __s0_903 = __p0_903;
  float16x4_t __s1_903 = __p1_903;
  float16x8_t __s2_903 = __p2_903;
  __ret_903 = vfmlsl_low_f16(__s0_903, __s1_903, (float16x4_t) {vgetq_lane_f16(__s2_903, __p3_903), vgetq_lane_f16(__s2_903, __p3_903), vgetq_lane_f16(__s2_903, __p3_903), vgetq_lane_f16(__s2_903, __p3_903)});
  __ret_903;
})

#define vmulh_lane_f16(__p0_905, __p1_905, __p2_905) __extension__ ({
  float16_t __ret_905;
  float16_t __s0_905 = __p0_905;
  float16x4_t __s1_905 = __p1_905;
  __ret_905 = __s0_905 * vget_lane_f16(__s1_905, __p2_905);
  __ret_905;
})

#define vmulh_laneq_f16(__p0_907, __p1_907, __p2_907) __extension__ ({
  float16_t __ret_907;
  float16_t __s0_907 = __p0_907;
  float16x8_t __s1_907 = __p1_907;
  __ret_907 = __s0_907 * vgetq_lane_f16(__s1_907, __p2_907);
  __ret_907;
})

__ai __attribute__((target("neon")))
uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2)

__ai __attribute__((target("neon")))
uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2)

__ai __attribute__((target("neon")))
uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2)

__ai __attribute__((target("neon")))
uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2)

__ai __attribute__((target("neon")))
uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2)

__ai __attribute__((target("neon")))
int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2)

__ai __attribute__((target("neon")))
uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2)

__ai __attribute__((target("neon")))
uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2)

__ai __attribute__((target("neon")))
int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2)

__ai __attribute__((target("neon")))
int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2)

#define vmulx_lane_f64(__p0_923, __p1_923, __p2_923) __extension__ ({
  float64x1_t __ret_923;
  float64x1_t __s0_923 = __p0_923;
  float64x1_t __s1_923 = __p1_923;
  float64_t __x_923 = vget_lane_f64(__s0_923, 0);
  float64_t __y_923 = vget_lane_f64(__s1_923, __p2_923);
  float64_t __z_923 = vmulxd_f64(__x_923, __y_923);
  __ret_923 = vset_lane_f64(__z_923, __s0_923, __p2_923);
  __ret_923;
})

#define vmulx_laneq_f64(__p0_924, __p1_924, __p2_924) __extension__ ({
  float64x1_t __ret_924;
  float64x1_t __s0_924 = __p0_924;
  float64x2_t __s1_924 = __p1_924;
  float64_t __x_924 = vget_lane_f64(__s0_924, 0);
  float64_t __y_924 = vgetq_lane_f64(__s1_924, __p2_924);
  float64_t __z_924 = vmulxd_f64(__x_924, __y_924);
  __ret_924 = vset_lane_f64(__z_924, __s0_924, 0);
  __ret_924;
})