/* SPDX-License-Identifier: MIT
 *
 * Permission is hereby granted, free of charge, to any person
 * obtaining a copy of this software and associated documentation
 * files (the "Software"), to deal in the Software without
 * restriction, including without limitation the rights to use, copy,
 * modify, merge, publish, distribute, sublicense, and/or sell copies
 * of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be
 * included in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 * Copyright:
 *   2020      Evan Nemerson <evan@nemerson.com>
 *   2020      Sean Maher <seanptmaher@gmail.com> (Copyright owned by Google, LLC)
 *   2023      Yi-Yen Chung <eric681@andestech.com> (Copyright owned by Andes Technology)
 *   2023      Yung-Cheng Su <eric20607@gapp.nthu.edu.tw> (Copyright owned by NTHU pllab)
 */

#if !defined(SIMDE_ARM_NEON_MUL_H)
#define SIMDE_ARM_NEON_MUL_H

#include "types.h"

#include "reinterpret.h"

HEDLEY_DIAGNOSTIC_PUSH
SIMDE_DISABLE_UNWANTED_DIAGNOSTICS
SIMDE_BEGIN_DECLS_

SIMDE_FUNCTION_ATTRIBUTES
simde_float16_t
simde_vmulh_f16(simde_float16_t a, simde_float16_t b) {
  #if defined(SIMDE_ARM_NEON_A32V8_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
    return vmulh_f16(a, b);
  #else
    simde_float32_t a_ = simde_float16_to_float32(a);
    simde_float32_t b_ = simde_float16_to_float32(b);

    return simde_float16_from_float32(a_ * b_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
  #undef vmulh_f16
  #define vmulh_f16(a, b) simde_vmulh_f16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_float16x4_t
simde_vmul_f16(simde_float16x4_t a, simde_float16x4_t b) {
  #if defined(SIMDE_ARM_NEON_A32V8_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
    return vmul_f16(a, b);
  #else
    simde_float16x4_private
      r_,
      a_ = simde_float16x4_to_private(a),
      b_ = simde_float16x4_to_private(b);

      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        simde_float32_t tmp_a_ = simde_float16_to_float32(a_.values[i]);
        simde_float32_t tmp_b_ = simde_float16_to_float32(b_.values[i]);
        r_.values[i] = simde_float16_from_float32(tmp_a_ * tmp_b_);
      }

    return simde_float16x4_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
  #undef vmul_f16
  #define vmul_f16(a, b) simde_vmul_f16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_float32x2_t
simde_vmul_f32(simde_float32x2_t a, simde_float32x2_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_f32(a, b);
  #else
    simde_float32x2_private
      r_,
      a_ = simde_float32x2_to_private(a),
      b_ = simde_float32x2_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vfmul_vv_f32m1(a_.sv64, b_.sv64, 2);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_float32x2_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_f32
  #define vmul_f32(a, b) simde_vmul_f32((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_float64x1_t
simde_vmul_f64(simde_float64x1_t a, simde_float64x1_t b) {
  #if defined(SIMDE_ARM_NEON_A64V8_NATIVE)
    return vmul_f64(a, b);
  #else
    simde_float64x1_private
      r_,
      a_ = simde_float64x1_to_private(a),
      b_ = simde_float64x1_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vfmul_vv_f64m1(a_.sv64, b_.sv64, 1);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_float64x1_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A64V8_ENABLE_NATIVE_ALIASES)
  #undef vmul_f64
  #define vmul_f64(a, b) simde_vmul_f64((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int8x8_t
simde_vmul_s8(simde_int8x8_t a, simde_int8x8_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_s8(a, b);
  #else
    simde_int8x8_private
      r_,
      a_ = simde_int8x8_to_private(a),
      b_ = simde_int8x8_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vmul_vv_i8m1(a_.sv64, b_.sv64, 8);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS) && !defined(SIMDE_BUG_GCC_100762)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_int8x8_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_s8
  #define vmul_s8(a, b) simde_vmul_s8((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int16x4_t
simde_vmul_s16(simde_int16x4_t a, simde_int16x4_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_s16(a, b);
  #else
    simde_int16x4_private
      r_,
      a_ = simde_int16x4_to_private(a),
      b_ = simde_int16x4_to_private(b);

    #if defined(SIMDE_X86_MMX_NATIVE)
      r_.m64 = _m_pmullw(a_.m64, b_.m64);
    #elif defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vmul_vv_i16m1(a_.sv64, b_.sv64, 4);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS) && !defined(SIMDE_BUG_GCC_100762)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_int16x4_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_s16
  #define vmul_s16(a, b) simde_vmul_s16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int32x2_t
simde_vmul_s32(simde_int32x2_t a, simde_int32x2_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_s32(a, b);
  #else
    simde_int32x2_private
      r_,
      a_ = simde_int32x2_to_private(a),
      b_ = simde_int32x2_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vmul_vv_i32m1(a_.sv64, b_.sv64, 2);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS) && !defined(SIMDE_BUG_GCC_100762)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_int32x2_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_s32
  #define vmul_s32(a, b) simde_vmul_s32((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int64x1_t
simde_x_vmul_s64(simde_int64x1_t a, simde_int64x1_t b) {
  simde_int64x1_private
    r_,
    a_ = simde_int64x1_to_private(a),
    b_ = simde_int64x1_to_private(b);

  #if defined(SIMDE_RISCV_V_NATIVE)
    r_.sv64 = __riscv_vmul_vv_i64m1(a_.sv64, b_.sv64, 1);
  #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
    r_.values = a_.values * b_.values;
  #else
    SIMDE_VECTORIZE
    for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
      r_.values[i] = a_.values[i] * b_.values[i];
    }
  #endif

  return simde_int64x1_from_private(r_);
}

SIMDE_FUNCTION_ATTRIBUTES
simde_uint8x8_t
simde_vmul_u8(simde_uint8x8_t a, simde_uint8x8_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_u8(a, b);
  #else
    simde_uint8x8_private
      r_,
      a_ = simde_uint8x8_to_private(a),
      b_ = simde_uint8x8_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vmul_vv_u8m1(a_.sv64, b_.sv64, 8);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS) && !defined(SIMDE_BUG_GCC_100762)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_uint8x8_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_u8
  #define vmul_u8(a, b) simde_vmul_u8((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_uint16x4_t
simde_vmul_u16(simde_uint16x4_t a, simde_uint16x4_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_u16(a, b);
  #else
    simde_uint16x4_private
      r_,
      a_ = simde_uint16x4_to_private(a),
      b_ = simde_uint16x4_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vmul_vv_u16m1(a_.sv64, b_.sv64, 4);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS) && !defined(SIMDE_BUG_GCC_100762)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_uint16x4_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_u16
  #define vmul_u16(a, b) simde_vmul_u16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_uint32x2_t
simde_vmul_u32(simde_uint32x2_t a, simde_uint32x2_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_u32(a, b);
  #else
    simde_uint32x2_private
      r_,
      a_ = simde_uint32x2_to_private(a),
      b_ = simde_uint32x2_to_private(b);

    #if defined(SIMDE_RISCV_V_NATIVE)
      r_.sv64 = __riscv_vmul_vv_u32m1(a_.sv64, b_.sv64, 2);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS) && !defined(SIMDE_BUG_GCC_100762)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_uint32x2_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_u32
  #define vmul_u32(a, b) simde_vmul_u32((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_uint64x1_t
simde_x_vmul_u64(simde_uint64x1_t a, simde_uint64x1_t b) {
  simde_uint64x1_private
    r_,
    a_ = simde_uint64x1_to_private(a),
    b_ = simde_uint64x1_to_private(b);

  #if defined(SIMDE_RISCV_V_NATIVE)
    r_.sv64 = __riscv_vmul_vv_u64m1(a_.sv64, b_.sv64, 1);
  #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
    r_.values = a_.values * b_.values;
  #else
    SIMDE_VECTORIZE
    for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
      r_.values[i] = a_.values[i] * b_.values[i];
    }
  #endif

  return simde_uint64x1_from_private(r_);
}

SIMDE_FUNCTION_ATTRIBUTES
simde_float16x8_t
simde_vmulq_f16(simde_float16x8_t a, simde_float16x8_t b) {
  #if defined(SIMDE_ARM_NEON_A32V8_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
    return vmulq_f16(a, b);
  #else
    simde_float16x8_private
      r_,
      a_ = simde_float16x8_to_private(a),
      b_ = simde_float16x8_to_private(b);

      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        simde_float32_t tmp_a_ = simde_float16_to_float32(a_.values[i]);
        simde_float32_t tmp_b_ = simde_float16_to_float32(b_.values[i]);
        r_.values[i] = simde_float16_from_float32(tmp_a_ * tmp_b_);
      }

    return simde_float16x8_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
  #undef vmulq_f16
  #define vmulq_f16(a, b) simde_vmulq_f16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_float32x4_t
simde_vmulq_f32(simde_float32x4_t a, simde_float32x4_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_f32(a, b);
  #else
    simde_float32x4_private
      r_,
      a_ = simde_float32x4_to_private(a),
      b_ = simde_float32x4_to_private(b);

    #if defined(SIMDE_X86_SSE_NATIVE)
      r_.m128 = _mm_mul_ps(a_.m128, b_.m128);
    #elif defined(SIMDE_WASM_SIMD128_NATIVE)
      r_.v128 = wasm_f32x4_mul(a_.v128, b_.v128);
    #elif defined(SIMDE_RISCV_V_NATIVE)
      r_.sv128 = __riscv_vfmul_vv_f32m1(a_.sv128, b_.sv128, 4);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_float32x4_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_f32
  #define vmulq_f32(a, b) simde_vmulq_f32((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_float64x2_t
simde_vmulq_f64(simde_float64x2_t a, simde_float64x2_t b) {
  #if defined(SIMDE_ARM_NEON_A64V8_NATIVE)
    return vmulq_f64(a, b);
  #else
    simde_float64x2_private
      r_,
      a_ = simde_float64x2_to_private(a),
      b_ = simde_float64x2_to_private(b);

    #if defined(SIMDE_X86_SSE2_NATIVE)
      r_.m128d = _mm_mul_pd(a_.m128d, b_.m128d);
    #elif defined(SIMDE_WASM_SIMD128_NATIVE)
      r_.v128 = wasm_f64x2_mul(a_.v128, b_.v128);
    #elif defined(SIMDE_RISCV_V_NATIVE)
      r_.sv128 = __riscv_vfmul_vv_f64m1(a_.sv128, b_.sv128, 2);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_float64x2_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A64V8_ENABLE_NATIVE_ALIASES)
  #undef vmulq_f64
  #define vmulq_f64(a, b) simde_vmulq_f64((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int8x16_t
simde_vmulq_s8(simde_int8x16_t a, simde_int8x16_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_s8(a, b);
  #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
    return vec_mul(a, b);
  #else
    simde_int8x16_private
      r_,
      a_ = simde_int8x16_to_private(a),
      b_ = simde_int8x16_to_private(b);

    #if defined(SIMDE_X86_SSE2_NATIVE)
      /* https://stackoverflow.com/a/29155682/501126 */
      const __m128i dst_even = _mm_mullo_epi16(a_.m128i, b_.m128i);
      r_.m128i =
        _mm_or_si128(
          _mm_slli_epi16(
            _mm_mullo_epi16(
              _mm_srli_epi16(a_.m128i, 8),
              _mm_srli_epi16(b_.m128i, 8)
            ),
            8
          ),
          #if defined(SIMDE_X86_AVX2_NATIVE)
            _mm_and_si128(dst_even, _mm_set1_epi16(0xFF))
          #else
            _mm_srli_epi16(
              _mm_slli_epi16(dst_even, 8),
              8
            )
          #endif
        );
    #elif defined(SIMDE_RISCV_V_NATIVE)
      r_.sv128 = __riscv_vmul_vv_i8m1(a_.sv128, b_.sv128, 16);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_int8x16_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_s8
  #define vmulq_s8(a, b) simde_vmulq_s8((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int16x8_t
simde_vmulq_s16(simde_int16x8_t a, simde_int16x8_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_s16(a, b);
  #else
    simde_int16x8_private
      r_,
      a_ = simde_int16x8_to_private(a),
      b_ = simde_int16x8_to_private(b);

    #if defined(SIMDE_X86_SSE2_NATIVE)
      r_.m128i = _mm_mullo_epi16(a_.m128i, b_.m128i);
    #elif defined(SIMDE_RISCV_V_NATIVE)
      r_.sv128 = __riscv_vmul_vv_i16m1(a_.sv128, b_.sv128, 8);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_int16x8_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_s16
  #define vmulq_s16(a, b) simde_vmulq_s16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int32x4_t
simde_vmulq_s32(simde_int32x4_t a, simde_int32x4_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_s32(a, b);
  #else
    simde_int32x4_private
      r_,
      a_ = simde_int32x4_to_private(a),
      b_ = simde_int32x4_to_private(b);

    #if defined(SIMDE_WASM_SIMD128_NATIVE)
      r_.v128 = wasm_i32x4_mul(a_.v128, b_.v128);
    #elif defined(SIMDE_RISCV_V_NATIVE)
      r_.sv128 = __riscv_vmul_vv_i32m1(a_.sv128, b_.sv128, 4);
    #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
      r_.values = a_.values * b_.values;
    #else
      SIMDE_VECTORIZE
      for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
        r_.values[i] = a_.values[i] * b_.values[i];
      }
    #endif

    return simde_int32x4_from_private(r_);
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_s32
  #define vmulq_s32(a, b) simde_vmulq_s32((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_int64x2_t
simde_x_vmulq_s64(simde_int64x2_t a, simde_int64x2_t b) {
  simde_int64x2_private
    r_,
    a_ = simde_int64x2_to_private(a),
    b_ = simde_int64x2_to_private(b);

  #if defined(SIMDE_WASM_SIMD128_NATIVE)
    r_.v128 = wasm_i64x2_mul(a_.v128, b_.v128);
  #elif defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512DQ_NATIVE)
    r_.m128i = _mm_mullo_epi64(a_.m128i, b_.m128i);
  #elif defined(SIMDE_RISCV_V_NATIVE)
    r_.sv128 = __riscv_vmul_vv_i64m1(a_.sv128, b_.sv128, 2);
  #elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
    r_.values = a_.values * b_.values;
  #else
    SIMDE_VECTORIZE
    for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
      r_.values[i] = a_.values[i] * b_.values[i];
    }
  #endif

  return simde_int64x2_from_private(r_);
}

SIMDE_FUNCTION_ATTRIBUTES
simde_uint8x16_t
simde_vmulq_u8(simde_uint8x16_t a, simde_uint8x16_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_u8(a, b);
  #elif defined(SIMDE_RISCV_V_NATIVE)
    simde_uint8x16_private
      r_,
      a_ = simde_uint8x16_to_private(a),
      b_ = simde_uint8x16_to_private(b);
    r_.sv128 = __riscv_vmul_vv_u8m1(a_.sv128, b_.sv128, 16);
    return simde_uint8x16_from_private(r_);
  #else
    return
      simde_vreinterpretq_u8_s8(
        simde_vmulq_s8(
          simde_vreinterpretq_s8_u8(a),
          simde_vreinterpretq_s8_u8(b)
        )
      );
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_u8
  #define vmulq_u8(a, b) simde_vmulq_u8((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_uint16x8_t
simde_vmulq_u16(simde_uint16x8_t a, simde_uint16x8_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_u16(a, b);
  #elif defined(SIMDE_RISCV_V_NATIVE)
    simde_uint16x8_private
      r_,
      a_ = simde_uint16x8_to_private(a),
      b_ = simde_uint16x8_to_private(b);
    r_.sv128 = __riscv_vmul_vv_u16m1(a_.sv128, b_.sv128, 8);
    return simde_uint16x8_from_private(r_);
  #else
    return
      simde_vreinterpretq_u16_s16(
        simde_vmulq_s16(
          simde_vreinterpretq_s16_u16(a),
          simde_vreinterpretq_s16_u16(b)
        )
      );
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_u16
  #define vmulq_u16(a, b) simde_vmulq_u16((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_uint32x4_t
simde_vmulq_u32(simde_uint32x4_t a, simde_uint32x4_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_u32(a, b);
  #elif defined(SIMDE_RISCV_V_NATIVE)
    simde_uint32x4_private
      r_,
      a_ = simde_uint32x4_to_private(a),
      b_ = simde_uint32x4_to_private(b);
    r_.sv128 = __riscv_vmul_vv_u32m1(a_.sv128, b_.sv128, 4);
    return simde_uint32x4_from_private(r_);
  #else
    return
      simde_vreinterpretq_u32_s32(
        simde_vmulq_s32(
          simde_vreinterpretq_s32_u32(a),
          simde_vreinterpretq_s32_u32(b)
        )
      );
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_u32
  #define vmulq_u32(a, b) simde_vmulq_u32((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_uint64x2_t
simde_x_vmulq_u64(simde_uint64x2_t a, simde_uint64x2_t b) {
  #if defined(SIMDE_RISCV_V_NATIVE)
    simde_uint64x2_private
      r_,
      a_ = simde_uint64x2_to_private(a),
      b_ = simde_uint64x2_to_private(b);
    r_.sv128 = __riscv_vmul_vv_u64m1(a_.sv128, b_.sv128, 2);
    return simde_uint64x2_from_private(r_);
  #else
    return
      simde_vreinterpretq_u64_s64(
        simde_x_vmulq_s64(
          simde_vreinterpretq_s64_u64(a),
          simde_vreinterpretq_s64_u64(b)
        )
      );
  #endif
}

SIMDE_FUNCTION_ATTRIBUTES
simde_poly8x8_t
simde_vmul_p8(simde_poly8x8_t a, simde_poly8x8_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmul_p8(a, b);
  #else
    simde_uint8x8_private
      r_,
      a_ = simde_uint8x8_to_private(simde_vreinterpret_u8_p8(a)),
      b_ = simde_uint8x8_to_private(simde_vreinterpret_u8_p8(b));

    SIMDE_VECTORIZE
    for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
      uint16_t extend_op2 = HEDLEY_STATIC_CAST(uint16_t, b_.values[i]);
      uint16_t result = 0;
      for(uint16_t j = 0; j < 8; ++j) {
        if (a_.values[i] & (1 << j)) {
          result = HEDLEY_STATIC_CAST(uint16_t, result ^ (extend_op2 << j));
        }
      }
      r_.values[i] = HEDLEY_STATIC_CAST(uint8_t, (result & (0xFF)));
    }

    return simde_vreinterpret_p8_u8(simde_uint8x8_from_private(r_));
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmul_p8
  #define vmul_p8(a, b) simde_vmul_p8((a), (b))
#endif

SIMDE_FUNCTION_ATTRIBUTES
simde_poly8x16_t
simde_vmulq_p8(simde_poly8x16_t a, simde_poly8x16_t b) {
  #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
    return vmulq_p8(a, b);
  #else
    simde_uint8x16_private
      r_,
      a_ = simde_uint8x16_to_private(simde_vreinterpretq_u8_p8(a)),
      b_ = simde_uint8x16_to_private(simde_vreinterpretq_u8_p8(b));

    SIMDE_VECTORIZE
    for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
      uint16_t extend_op2 = HEDLEY_STATIC_CAST(uint16_t, b_.values[i]);
      uint16_t result = 0;
      for(uint16_t j = 0; j < 8; ++j) {
        if (a_.values[i] & (1 << j)) {
          result = HEDLEY_STATIC_CAST(uint16_t, result ^ (extend_op2 << j));
        }
      }
      r_.values[i] = HEDLEY_STATIC_CAST(uint8_t, (result & (0xFF)));
    }

    return simde_vreinterpretq_p8_u8(simde_uint8x16_from_private(r_));
  #endif
}
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
  #undef vmulq_p8
  #define vmulq_p8(a, b) simde_vmulq_p8((a), (b))
#endif

SIMDE_END_DECLS_
HEDLEY_DIAGNOSTIC_POP

#endif /* !defined(SIMDE_ARM_NEON_MUL_H) */