// Copyright 2019 Google LLC
// SPDX-License-Identifier: Apache-2.0
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//      http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

#include <stddef.h>
#include <stdint.h>

#undef HWY_TARGET_INCLUDE
#define HWY_TARGET_INCLUDE "tests/arithmetic_test.cc"
#include "hwy/foreach_target.h"  // IWYU pragma: keep
#include "hwy/highway.h"
#include "hwy/nanobenchmark.h"
#include "hwy/tests/test_util-inl.h"

HWY_BEFORE_NAMESPACE();
namespace hwy {
namespace HWY_NAMESPACE {

struct TestPlusMinus {
  template <class T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const auto v2 = Iota(d, 2);
    const auto v3 = Iota(d, 3);
    const auto v4 = Iota(d, 4);

    const size_t N = Lanes(d);
    auto lanes = AllocateAligned<T>(N);
    HWY_ASSERT(lanes);
    for (size_t i = 0; i < N; ++i) {
      lanes[i] = ConvertScalarTo<T>((2 + i) + (3 + i));
    }
    HWY_ASSERT_VEC_EQ(d, lanes.get(), Add(v2, v3));
    HWY_ASSERT_VEC_EQ(d, Set(d, ConvertScalarTo<T>(2)), Sub(v4, v2));

    for (size_t i = 0; i < N; ++i) {
      lanes[i] = ConvertScalarTo<T>((2 + i) + (4 + i));
    }
    auto sum = v2;
    sum = Add(sum, v4);  // sum == 6,8..
    HWY_ASSERT_VEC_EQ(d, Load(d, lanes.get()), sum);

    sum = Sub(sum, v4);
    HWY_ASSERT_VEC_EQ(d, v2, sum);
  }
};

struct TestPlusMinusOverflow {
  template <class T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const auto v1 = Iota(d, 1);
    const auto vMax = Iota(d, LimitsMax<T>());
    const auto vMin = Iota(d, LimitsMin<T>());

    // Check that no UB triggered.
    // "assert" here is formal - to avoid compiler dropping calculations
    HWY_ASSERT_VEC_EQ(d, Add(v1, vMax), Add(vMax, v1));
    HWY_ASSERT_VEC_EQ(d, Add(vMax, vMax), Add(vMax, vMax));
    HWY_ASSERT_VEC_EQ(d, Sub(vMin, v1), Sub(vMin, v1));
    HWY_ASSERT_VEC_EQ(d, Sub(vMin, vMax), Sub(vMin, vMax));
  }
};

HWY_NOINLINE void TestAllPlusMinus() {
  ForAllTypes(ForPartialVectors<TestPlusMinus>());
  ForIntegerTypes(ForPartialVectors<TestPlusMinusOverflow>());
}

struct TestAddSub {
  template <class T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const auto v2 = Iota(d, 2);
    const auto v4 = Iota(d, 4);

    const size_t N = Lanes(d);
    auto lanes = AllocateAligned<T>(N);
    HWY_ASSERT(lanes);
    for (size_t i = 0; i < N; ++i) {
      lanes[i] = ConvertScalarTo<T>(((i & 1) == 0) ? 2 : ((4 + i) + (2 + i)));
    }
    HWY_ASSERT_VEC_EQ(d, lanes.get(), AddSub(v4, v2));

    for (size_t i = 0; i < N; ++i) {
      if ((i & 1) == 0) {
        lanes[i] = ConvertScalarTo<T>(-2);
      }
    }
    HWY_ASSERT_VEC_EQ(d, lanes.get(), AddSub(v2, v4));
  }
};

HWY_NOINLINE void TestAllAddSub() {
  ForAllTypes(ForPartialVectors<TestAddSub>());
}

struct TestUnsignedSaturatingArithmetic {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const auto v0 = Zero(d);
    const auto vi = Iota(d, 1);
    const auto vm = Set(d, LimitsMax<T>());

    HWY_ASSERT_VEC_EQ(d, Add(v0, v0), SaturatedAdd(v0, v0));
    HWY_ASSERT_VEC_EQ(d, Add(v0, vi), SaturatedAdd(v0, vi));
    HWY_ASSERT_VEC_EQ(d, Add(v0, vm), SaturatedAdd(v0, vm));
    HWY_ASSERT_VEC_EQ(d, vm, SaturatedAdd(vi, vm));
    HWY_ASSERT_VEC_EQ(d, vm, SaturatedAdd(vm, vm));

    HWY_ASSERT_VEC_EQ(d, v0, SaturatedSub(v0, v0));
    HWY_ASSERT_VEC_EQ(d, v0, SaturatedSub(v0, vi));
    HWY_ASSERT_VEC_EQ(d, v0, SaturatedSub(vi, vi));
    HWY_ASSERT_VEC_EQ(d, v0, SaturatedSub(vi, vm));
    HWY_ASSERT_VEC_EQ(d, Sub(vm, vi), SaturatedSub(vm, vi));
  }
};

struct TestSignedSaturatingArithmetic {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const Vec<D> v0 = Zero(d);
    const Vec<D> vpm = Set(d, LimitsMax<T>());
    const Vec<D> vi = PositiveIota(d);
    const Vec<D> vn = Sub(v0, vi);
    const Vec<D> vnm = Set(d, LimitsMin<T>());
    HWY_ASSERT_MASK_EQ(d, MaskTrue(d), Gt(vi, v0));
    HWY_ASSERT_MASK_EQ(d, MaskTrue(d), Lt(vn, v0));

    HWY_ASSERT_VEC_EQ(d, v0, SaturatedAdd(v0, v0));
    HWY_ASSERT_VEC_EQ(d, vi, SaturatedAdd(v0, vi));
    HWY_ASSERT_VEC_EQ(d, vpm, SaturatedAdd(v0, vpm));
    HWY_ASSERT_VEC_EQ(d, vpm, SaturatedAdd(vi, vpm));
    HWY_ASSERT_VEC_EQ(d, vpm, SaturatedAdd(vpm, vpm));

    HWY_ASSERT_VEC_EQ(d, v0, SaturatedSub(v0, v0));
    HWY_ASSERT_VEC_EQ(d, Sub(v0, vi), SaturatedSub(v0, vi));
    HWY_ASSERT_VEC_EQ(d, vn, SaturatedSub(vn, v0));
    HWY_ASSERT_VEC_EQ(d, vnm, SaturatedSub(vnm, vi));
    HWY_ASSERT_VEC_EQ(d, vnm, SaturatedSub(vnm, vpm));
  }
};

struct TestSaturatingArithmeticOverflow {
  template <class T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const auto v1 = Iota(d, 1);
    const auto vMax = Iota(d, LimitsMax<T>());
    const auto vMin = Iota(d, LimitsMin<T>());

    // Check that no UB triggered.
    // "assert" here is formal - to avoid compiler dropping calculations
    HWY_ASSERT_VEC_EQ(d, SaturatedAdd(v1, vMax), SaturatedAdd(vMax, v1));
    HWY_ASSERT_VEC_EQ(d, SaturatedAdd(vMax, vMax), SaturatedAdd(vMax, vMax));
    HWY_ASSERT_VEC_EQ(d, SaturatedAdd(vMin, vMax), SaturatedAdd(vMin, vMax));
    HWY_ASSERT_VEC_EQ(d, SaturatedAdd(vMin, vMin), SaturatedAdd(vMin, vMin));
    HWY_ASSERT_VEC_EQ(d, SaturatedSub(vMin, v1), SaturatedSub(vMin, v1));
    HWY_ASSERT_VEC_EQ(d, SaturatedSub(vMin, vMax), SaturatedSub(vMin, vMax));
    HWY_ASSERT_VEC_EQ(d, SaturatedSub(vMax, vMin), SaturatedSub(vMax, vMin));
    HWY_ASSERT_VEC_EQ(d, SaturatedSub(vMin, vMin), SaturatedSub(vMin, vMin));
  }
};

HWY_NOINLINE void TestAllSaturatingArithmetic() {
  ForUnsignedTypes(ForPartialVectors<TestUnsignedSaturatingArithmetic>());
  ForSignedTypes(ForPartialVectors<TestSignedSaturatingArithmetic>());
  ForIntegerTypes(ForPartialVectors<TestSaturatingArithmeticOverflow>());
}

struct TestAverage {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const Vec<D> v0 = Zero(d);
    const Vec<D> v1 = Set(d, static_cast<T>(1));
    const Vec<D> v2 = Set(d, static_cast<T>(2));

    HWY_ASSERT_VEC_EQ(d, v0, AverageRound(v0, v0));
    HWY_ASSERT_VEC_EQ(d, v1, AverageRound(v0, v1));
    HWY_ASSERT_VEC_EQ(d, v1, AverageRound(v1, v1));
    HWY_ASSERT_VEC_EQ(d, v2, AverageRound(v1, v2));
    HWY_ASSERT_VEC_EQ(d, v2, AverageRound(v2, v2));
  }
};

HWY_NOINLINE void TestAllAverage() {
  const ForPartialVectors<TestAverage> test;
  test(uint8_t());
  test(uint16_t());
}

struct TestAbs {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const Vec<D> v0 = Zero(d);
    const Vec<D> vp1 = Set(d, static_cast<T>(1));
    const Vec<D> vn1 = Set(d, static_cast<T>(-1));
    const Vec<D> vpm = Set(d, LimitsMax<T>());
    const Vec<D> vnm = Set(d, LimitsMin<T>());

    HWY_ASSERT_VEC_EQ(d, v0, Abs(v0));
    HWY_ASSERT_VEC_EQ(d, vp1, Abs(vp1));
    HWY_ASSERT_VEC_EQ(d, vp1, Abs(vn1));
    HWY_ASSERT_VEC_EQ(d, vpm, Abs(vpm));
    HWY_ASSERT_VEC_EQ(d, vnm, Abs(vnm));
  }
};

struct TestFloatAbs {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const Vec<D> v0 = Zero(d);
    const Vec<D> vp1 = Set(d, ConvertScalarTo<T>(1));
    const Vec<D> vn1 = Set(d, ConvertScalarTo<T>(-1));
    const Vec<D> vp2 = Set(d, ConvertScalarTo<T>(0.01));
    const Vec<D> vn2 = Set(d, ConvertScalarTo<T>(-0.01));

    HWY_ASSERT_VEC_EQ(d, v0, Abs(v0));
    HWY_ASSERT_VEC_EQ(d, vp1, Abs(vp1));
    HWY_ASSERT_VEC_EQ(d, vp1, Abs(vn1));
    HWY_ASSERT_VEC_EQ(d, vp2, Abs(vp2));
    HWY_ASSERT_VEC_EQ(d, vp2, Abs(vn2));
  }
};

HWY_NOINLINE void TestAllAbs() {
  ForSignedTypes(ForPartialVectors<TestAbs>());
  ForFloatTypes(ForPartialVectors<TestFloatAbs>());
}

struct TestSaturatedAbs {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const Vec<D> v0 = Zero(d);
    const Vec<D> vp1 = Set(d, static_cast<T>(1));
    const Vec<D> vn1 = Set(d, static_cast<T>(-1));
    const Vec<D> vpm = Set(d, LimitsMax<T>());
    const Vec<D> vnm = Set(d, LimitsMin<T>());

    HWY_ASSERT_VEC_EQ(d, v0, SaturatedAbs(v0));
    HWY_ASSERT_VEC_EQ(d, vp1, SaturatedAbs(vp1));
    HWY_ASSERT_VEC_EQ(d, vp1, SaturatedAbs(vn1));
    HWY_ASSERT_VEC_EQ(d, vpm, SaturatedAbs(vpm));
    HWY_ASSERT_VEC_EQ(d, vpm, SaturatedAbs(vnm));
  }
};

HWY_NOINLINE void TestAllSaturatedAbs() {
  ForSignedTypes(ForPartialVectors<TestSaturatedAbs>());
}

struct TestIntegerNeg {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const RebindToUnsigned<D> du;
    using TU = TFromD<decltype(du)>;
    const Vec<D> v0 = Zero(d);
    const Vec<D> v1 = BitCast(d, Set(du, TU{1}));
    const Vec<D> vp = BitCast(d, Set(du, TU{3}));
    const Vec<D> vn = Add(Not(vp), v1);  // 2's complement
    HWY_ASSERT_VEC_EQ(d, v0, Neg(v0));
    HWY_ASSERT_VEC_EQ(d, vp, Neg(vn));
    HWY_ASSERT_VEC_EQ(d, vn, Neg(vp));
  }
};

struct TestFloatNeg {
  // Must be inlined on aarch64 for bf16, else clang crashes.
  template <typename T, class D>
  HWY_INLINE void operator()(T /*unused*/, D d) {
    const RebindToUnsigned<D> du;
    using TU = TFromD<decltype(du)>;
    // 1.25 in binary16.
    const Vec<D> vp =
        BitCast(d, Set(du, static_cast<TU>(Unpredictable1() * 0x3D00)));
    // Flip sign bit in MSB
    const Vec<D> vn = BitCast(d, Xor(BitCast(du, vp), SignBit(du)));
    // Do not check negative zero - we do not yet have proper bfloat16_t Eq().
    HWY_ASSERT_VEC_EQ(du, BitCast(du, vp), BitCast(du, Neg(vn)));
    HWY_ASSERT_VEC_EQ(du, BitCast(du, vn), BitCast(du, Neg(vp)));
  }
};

struct TestNegOverflow {
  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const auto vn = Set(d, LimitsMin<T>());
    const auto vp = Set(d, LimitsMax<T>());
    HWY_ASSERT_VEC_EQ(d, Neg(vn), Neg(vn));
    HWY_ASSERT_VEC_EQ(d, Neg(vp), Neg(vp));
  }
};

HWY_NOINLINE void TestAllNeg() {
  ForFloatTypes(ForPartialVectors<TestFloatNeg>());
  // Always supported, even if !HWY_HAVE_FLOAT16.
  ForPartialVectors<TestFloatNeg>()(float16_t());

  ForSignedTypes(ForPartialVectors<TestIntegerNeg>());

  ForSignedTypes(ForPartialVectors<TestNegOverflow>());
}

struct TestSaturatedNeg {
  template <class D>
  static HWY_NOINLINE void VerifySatNegOverflow(D d) {
    using T = TFromD<D>;
    HWY_ASSERT_VEC_EQ(d, Set(d, LimitsMax<T>()),
                      SaturatedNeg(Set(d, LimitsMin<T>())));
  }

  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    VerifySatNegOverflow(d);

    const RebindToUnsigned<D> du;
    using TU = TFromD<decltype(du)>;
    const Vec<D> v0 = Zero(d);
    const Vec<D> v1 = BitCast(d, Set(du, TU{1}));
    const Vec<D> vp = BitCast(d, Set(du, TU{3}));
    const Vec<D> vn = Add(Not(vp), v1);  // 2's complement
    HWY_ASSERT_VEC_EQ(d, v0, SaturatedNeg(v0));
    HWY_ASSERT_VEC_EQ(d, vp, SaturatedNeg(vn));
    HWY_ASSERT_VEC_EQ(d, vn, SaturatedNeg(vp));
  }
};

HWY_NOINLINE void TestAllSaturatedNeg() {
  ForSignedTypes(ForPartialVectors<TestSaturatedNeg>());
}

struct TestIntegerAbsDiff {
  template <typename T, HWY_IF_T_SIZE_ONE_OF(T, (1 << 1) | (1 << 2) | (1 << 4))>
  static inline T ScalarAbsDiff(T a, T b) {
    using TW = MakeSigned<MakeWide<T>>;
    const TW diff = static_cast<TW>(static_cast<TW>(a) - static_cast<TW>(b));
    return static_cast<T>((diff >= 0) ? diff : -diff);
  }
  template <typename T, HWY_IF_T_SIZE(T, 8)>
  static inline T ScalarAbsDiff(T a, T b) {
    if (a >= b) {
      return static_cast<T>(static_cast<uint64_t>(a) -
                            static_cast<uint64_t>(b));
    } else {
      return static_cast<T>(static_cast<uint64_t>(b) -
                            static_cast<uint64_t>(a));
    }
  }

  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    const size_t N = Lanes(d);
    auto in_lanes_a = AllocateAligned<T>(N);
    auto in_lanes_b = AllocateAligned<T>(N);
    auto out_lanes = AllocateAligned<T>(N);
    constexpr size_t shift_amt_mask = sizeof(T) * 8 - 1;
    for (size_t i = 0; i < N; ++i) {
      // Need to mask out shift_amt as i can be greater than or equal to
      // the number of bits in T if T is int8_t, uint8_t, int16_t, or uint16_t.
      const auto shift_amt = i & shift_amt_mask;
      in_lanes_a[i] =
          static_cast<T>((static_cast<uint64_t>(i) ^ 1u) << shift_amt);
      in_lanes_b[i] = static_cast<T>(static_cast<uint64_t>(i) << shift_amt);
      out_lanes[i] = ScalarAbsDiff(in_lanes_a[i], in_lanes_b[i]);
    }
    const auto a = Load(d, in_lanes_a.get());
    const auto b = Load(d, in_lanes_b.get());
    const auto expected = Load(d, out_lanes.get());
    HWY_ASSERT_VEC_EQ(d, expected, AbsDiff(a, b));
    HWY_ASSERT_VEC_EQ(d, expected, AbsDiff(b, a));
  }
};

HWY_NOINLINE void TestAllIntegerAbsDiff() {
  ForPartialVectors<TestIntegerAbsDiff>()(int8_t());
  ForPartialVectors<TestIntegerAbsDiff>()(uint8_t());
  ForPartialVectors<TestIntegerAbsDiff>()(int16_t());
  ForPartialVectors<TestIntegerAbsDiff>()(uint16_t());
  ForPartialVectors<TestIntegerAbsDiff>()(int32_t());
  ForPartialVectors<TestIntegerAbsDiff>()(uint32_t());
#if HWY_HAVE_INTEGER64
  ForPartialVectors<TestIntegerAbsDiff>()(int64_t());
  ForPartialVectors<TestIntegerAbsDiff>()(uint64_t());
#endif
}

struct TestIntegerDiv {
  template <class D>
  static HWY_NOINLINE void DoTestIntegerDiv(D d, const VecArg<VFromD<D>> a,
                                            const VecArg<VFromD<D>> b) {
    using T = TFromD<D>;

    const size_t N = Lanes(d);
    auto a_lanes = AllocateAligned<T>(N);
    auto b_lanes = AllocateAligned<T>(N);
    auto expected = AllocateAligned<T>(N);
    auto expected_even = AllocateAligned<T>(N);
    auto expected_odd = AllocateAligned<T>(N);
    HWY_ASSERT(a_lanes && b_lanes && expected && expected_even && expected_odd);

    Store(a, d, a_lanes.get());
    Store(b, d, b_lanes.get());

    for (size_t i = 0; i < N; i++) {
      expected[i] = static_cast<T>(a_lanes[i] / b_lanes[i]);
      if ((i & 1) == 0) {
        expected_even[i] = expected[i];
        expected_odd[i] = static_cast<T>(0);
      } else {
        expected_even[i] = static_cast<T>(0);
        expected_odd[i] = expected[i];
      }
    }

    HWY_ASSERT_VEC_EQ(d, expected.get(), Div(a, b));

    const auto vmin = Set(d, LimitsMin<T>());
    const auto zero = Zero(d);
    const auto all_ones = Set(d, static_cast<T>(-1));

    HWY_ASSERT_VEC_EQ(d, expected_even.get(),
                      OddEven(zero, Div(a, OddEven(zero, b))));
    HWY_ASSERT_VEC_EQ(d, expected_odd.get(),
                      OddEven(Div(a, OddEven(b, zero)), zero));

    HWY_ASSERT_VEC_EQ(
        d, expected_even.get(),
        OddEven(zero, Div(OddEven(vmin, a), OddEven(all_ones, b))));
    HWY_ASSERT_VEC_EQ(
        d, expected_odd.get(),
        OddEven(Div(OddEven(a, vmin), OddEven(b, all_ones)), zero));
  }

  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    using TU = MakeUnsigned<T>;
    using TI = MakeSigned<T>;

    const size_t N = Lanes(d);

#if HWY_TARGET <= HWY_AVX3 && HWY_IS_MSAN
    // Workaround for MSAN bug on AVX3
    if (sizeof(T) <= 2 && N >= 16) {
      return;
    }
#endif

    const auto vmin = Set(d, LimitsMin<T>());
    const auto vmax = Set(d, LimitsMax<T>());
    const auto v2 = Set(d, static_cast<T>(Unpredictable1() + 1));
    const auto v3 = Set(d, static_cast<T>(Unpredictable1() + 2));

    HWY_ASSERT_VEC_EQ(d, Set(d, static_cast<T>(LimitsMin<T>() / 2)),
                      Div(vmin, v2));
    HWY_ASSERT_VEC_EQ(d, Set(d, static_cast<T>(LimitsMin<T>() / 3)),
                      Div(vmin, v3));
    HWY_ASSERT_VEC_EQ(d, Set(d, static_cast<T>(LimitsMax<T>() / 2)),
                      Div(vmax, v2));
    HWY_ASSERT_VEC_EQ(d, Set(d, static_cast<T>(LimitsMax<T>() / 3)),
                      Div(vmax, v3));

    auto in1 = AllocateAligned<T>(N);
    auto in2 = AllocateAligned<T>(N);
    HWY_ASSERT(in1 && in2);

    const RebindToSigned<decltype(d)> di;

    // Random inputs in each lane
    RandomState rng;
    for (size_t rep = 0; rep < AdjustedReps(1000); ++rep) {
      for (size_t i = 0; i < N; ++i) {
        const T rnd_a0 = static_cast<T>(Random64(&rng) &
                                        static_cast<uint64_t>(LimitsMax<TU>()));
        const T rnd_b0 = static_cast<T>(Random64(&rng) &
                                        static_cast<uint64_t>(LimitsMax<TI>()));

        const T rnd_b = static_cast<T>(rnd_b0 | static_cast<T>(rnd_b0 == 0));
        const T rnd_a = static_cast<T>(
            rnd_a0 + static_cast<T>(IsSigned<T>() && rnd_a0 == LimitsMin<T>() &&
                                    ScalarAbs(rnd_b) == static_cast<T>(1)));

        in1[i] = rnd_a;
        in2[i] = rnd_b;
      }

      const auto a = Load(d, in1.get());
      const auto b = Load(d, in2.get());

      const auto neg_a = BitCast(d, Neg(BitCast(di, a)));
      const auto neg_b = BitCast(d, Neg(BitCast(di, b)));

      DoTestIntegerDiv(d, a, b);
      DoTestIntegerDiv(d, a, neg_b);
      DoTestIntegerDiv(d, neg_a, b);
      DoTestIntegerDiv(d, neg_a, neg_b);
    }
  }
};

HWY_NOINLINE void TestAllIntegerDiv() {
  ForIntegerTypes(ForPartialVectors<TestIntegerDiv>());
}

struct TestIntegerMod {
  template <class D>
  static HWY_NOINLINE void DoTestIntegerMod(D d, const VecArg<VFromD<D>> a,
                                            const VecArg<VFromD<D>> b) {
    using T = TFromD<D>;

    const size_t N = Lanes(d);

#if HWY_TARGET <= HWY_AVX3 && HWY_IS_MSAN
    // Workaround for MSAN bug on AVX3
    if (sizeof(T) <= 2 && N >= 16) {
      return;
    }
#endif

    auto a_lanes = AllocateAligned<T>(N);
    auto b_lanes = AllocateAligned<T>(N);
    auto expected = AllocateAligned<T>(N);
    auto expected_even = AllocateAligned<T>(N);
    auto expected_odd = AllocateAligned<T>(N);
    HWY_ASSERT(a_lanes && b_lanes && expected && expected_even && expected_odd);

    Store(a, d, a_lanes.get());
    Store(b, d, b_lanes.get());

    for (size_t i = 0; i < N; i++) {
      expected[i] = static_cast<T>(a_lanes[i] % b_lanes[i]);
      if ((i & 1) == 0) {
        expected_even[i] = expected[i];
        expected_odd[i] = static_cast<T>(0);
      } else {
        expected_even[i] = static_cast<T>(0);
        expected_odd[i] = expected[i];
      }
    }

    HWY_ASSERT_VEC_EQ(d, expected.get(), Mod(a, b));

    const auto vmin = Set(d, LimitsMin<T>());
    const auto zero = Zero(d);
    const auto all_ones = Set(d, static_cast<T>(-1));

    HWY_ASSERT_VEC_EQ(d, expected_even.get(),
                      OddEven(zero, Mod(a, OddEven(zero, b))));
    HWY_ASSERT_VEC_EQ(d, expected_odd.get(),
                      OddEven(Mod(a, OddEven(b, zero)), zero));

    HWY_ASSERT_VEC_EQ(
        d, expected_even.get(),
        OddEven(zero, Mod(OddEven(vmin, a), OddEven(all_ones, b))));
    HWY_ASSERT_VEC_EQ(
        d, expected_odd.get(),
        OddEven(Mod(OddEven(a, vmin), OddEven(b, all_ones)), zero));
  }

  template <typename T, class D>
  HWY_NOINLINE void operator()(T /*unused*/, D d) {
    using TU = MakeUnsigned<T>;
    using TI = MakeSigned<T>;

    const size_t N = Lanes(d);
    auto in1 = AllocateAligned<T>(N);
    auto in2 = AllocateAligned<T>(N);
    HWY_ASSERT(in1 && in2);

    const RebindToSigned<decltype(d)> di;

    // Random inputs in each lane
    RandomState rng;
    for (size_t rep = 0; rep < AdjustedReps(1000); ++rep) {
      for (size_t i = 0; i < N; ++i) {
        const T rnd_a0 = static_cast<T>(Random64(&rng) &
                                        static_cast<uint64_t>(LimitsMax<TU>()));
        const T rnd_b0 = static_cast<T>(Random64(&rng) &
                                        static_cast<uint64_t>(LimitsMax<TI>()));

        const T rnd_b = static_cast<T>(rnd_b0 | static_cast<T>(rnd_b0 == 0));
        const T rnd_a = static_cast<T>(
            rnd_a0 + static_cast<T>(IsSigned<T>() && rnd_a0 == LimitsMin<T>() &&
                                    ScalarAbs(rnd_b) == static_cast<T>(1)));

        in1[i] = rnd_a;
        in2[i] = rnd_b;
      }

      const auto a = Load(d, in1.get());
      const auto b = Load(d, in2.get());

      const auto neg_a = BitCast(d, Neg(BitCast(di, a)));
      const auto neg_b = BitCast(d, Neg(BitCast(di, b)));

      DoTestIntegerMod(d, a, b);
      DoTestIntegerMod(d, a, neg_b);
      DoTestIntegerMod(d, neg_a, b);
      DoTestIntegerMod(d, neg_a, neg_b);
    }
  }
};

HWY_NOINLINE void TestAllIntegerMod() {
  ForIntegerTypes(ForPartialVectors<TestIntegerMod>());
}

// NOLINTNEXTLINE(google-readability-namespace-comments)
}  // namespace HWY_NAMESPACE
}  // namespace hwy
HWY_AFTER_NAMESPACE();

#if HWY_ONCE

namespace hwy {
HWY_BEFORE_TEST(HwyArithmeticTest);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllPlusMinus);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllAddSub);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllSaturatingArithmetic);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllAverage);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllAbs);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllSaturatedAbs);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllNeg);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllSaturatedNeg);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllIntegerAbsDiff);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllIntegerDiv);
HWY_EXPORT_AND_TEST_P(HwyArithmeticTest, TestAllIntegerMod);
}  // namespace hwy

#endif