// MIT License
//
// Copyright (c) 2017 Advanced Micro Devices, Inc. All rights reserved.
//
// 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.

#include <algorithm>
#include <functional>
#include <iostream>
#include <type_traits>
#include <vector>
#include <utility>

// Google Test
#include <gtest/gtest.h>
// rocPRIM API
#include <rocprim/rocprim.hpp>

#include "test_utils.hpp"
#include "test_utils_types.hpp"

#define HIP_CHECK(error) ASSERT_EQ(static_cast<hipError_t>(error), hipSuccess)

namespace rp = rocprim;

template<class Params>
class RocprimBlockRadixSort : public ::testing::Test {
public:
    using params = Params;
};

static constexpr size_t n_sizes = 12;
static constexpr unsigned int items_radix[n_sizes] = {
    1, 2, 3, 1, 2, 3, 1, 2, 3, 1, 2, 3
};
static constexpr bool desc_radix[n_sizes] = {
    false, false, false, false, false, false, true, true, true, true, true, true
};
static constexpr bool striped_radix[n_sizes] = {
    false, false, false, true, true, true, false, false, false, true, true, true
};
static constexpr unsigned int start_radix[n_sizes] = {
    0, 0, 0, 3, 4, 8, 0, 0, 0, 3, 4, 8
};
static constexpr unsigned int end_radix[n_sizes] = {
    0, 0, 0, 10, 11, 12, 0, 0, 0, 10, 11, 12
};

TYPED_TEST_CASE(RocprimBlockRadixSort, BlockParams);

template<class Key, bool Descending, unsigned int StartBit, unsigned int EndBit>
struct key_comparator
{
    static_assert(rp::is_unsigned<Key>::value, "Test supports start and end bits only for unsigned integers");

    bool operator()(const Key& lhs, const Key& rhs)
    {
        auto mask = (1ull << (EndBit - StartBit)) - 1;
        auto l = (static_cast<unsigned long long>(lhs) >> StartBit) & mask;
        auto r = (static_cast<unsigned long long>(rhs) >> StartBit) & mask;
        return Descending ? (r < l) : (l < r);
    }
};

template<class Key, bool Descending>
struct key_comparator<Key, Descending, 0, sizeof(Key) * 8>
{
    bool operator()(const Key& lhs, const Key& rhs)
    {
        return Descending ? (rhs < lhs) : (lhs < rhs);
    }
};

template<bool Descending>
struct key_comparator<rp::half, Descending, 0, sizeof(rp::half) * 8>
{
    bool operator()(const rp::half& lhs, const rp::half& rhs)
    {
        // HIP's half doesn't have __host__ comparison operators, use floats instead
        return key_comparator<float, Descending, 0, sizeof(float) * 8>()(lhs, rhs);
    }
};

template<class Key, class Value, bool Descending, unsigned int StartBit, unsigned int EndBit>
struct key_value_comparator
{
    bool operator()(const std::pair<Key, Value>& lhs, const std::pair<Key, Value>& rhs)
    {
        return key_comparator<Key, Descending, StartBit, EndBit>()(lhs.first, rhs.first);
    }
};

template<
    unsigned int BlockSize,
    unsigned int ItemsPerThread,
    class key_type
>
__global__
void sort_key_kernel(
    key_type* device_keys_output,
    bool to_striped,
    bool descending,
    unsigned int start_bit,
    unsigned int end_bit)
{
    constexpr unsigned int items_per_block = BlockSize * ItemsPerThread;
    const unsigned int lid = hipThreadIdx_x;
    const unsigned int block_offset = hipBlockIdx_x * items_per_block;

    key_type keys[ItemsPerThread];
    rp::block_load_direct_blocked(lid, device_keys_output + block_offset, keys);

    rp::block_radix_sort<key_type, BlockSize, ItemsPerThread> bsort;

    if(to_striped)
    {
        if(descending)
            bsort.sort_desc_to_striped(keys, start_bit, end_bit);
        else
            bsort.sort_to_striped(keys, start_bit, end_bit);

        rp::block_store_direct_striped<BlockSize>(lid, device_keys_output + block_offset, keys);
    }
    else
    {
        if(descending)
            bsort.sort_desc(keys, start_bit, end_bit);
        else
            bsort.sort(keys, start_bit, end_bit);

        rp::block_store_direct_blocked(lid, device_keys_output + block_offset, keys);
    }
}

template<
    unsigned int BlockSize,
    unsigned int ItemsPerThread,
    class key_type,
    class value_type
>
__global__
void sort_key_value_kernel(
    key_type* device_keys_output,
    value_type* device_values_output,
    bool to_striped,
    bool descending,
    unsigned int start_bit,
    unsigned int end_bit)
{
    constexpr unsigned int items_per_block = BlockSize * ItemsPerThread;
    const unsigned int lid = hipThreadIdx_x;
    const unsigned int block_offset = hipBlockIdx_x * items_per_block;

    key_type keys[ItemsPerThread];
    value_type values[ItemsPerThread];
    rp::block_load_direct_blocked(lid, device_keys_output + block_offset, keys);
    rp::block_load_direct_blocked(lid, device_values_output + block_offset, values);

    rp::block_radix_sort<key_type, BlockSize, ItemsPerThread, value_type> bsort;
    if(to_striped)
    {
        if(descending)
            bsort.sort_desc_to_striped(keys, values, start_bit, end_bit);
        else
            bsort.sort_to_striped(keys, values, start_bit, end_bit);

        rp::block_store_direct_striped<BlockSize>(lid, device_keys_output + block_offset, keys);
        rp::block_store_direct_striped<BlockSize>(lid, device_values_output + block_offset, values);
    }
    else
    {
        if(descending)
            bsort.sort_desc(keys, values, start_bit, end_bit);
        else
            bsort.sort(keys, values, start_bit, end_bit);

        rp::block_store_direct_blocked(lid, device_keys_output + block_offset, keys);
        rp::block_store_direct_blocked(lid, device_values_output + block_offset, values);
    }
}

// Test for radix sort
template<
    class Key,
    class Value,
    unsigned int Method,
    unsigned int BlockSize,
    unsigned int ItemsPerThread,
    bool Descending = false,
    bool ToStriped = false,
    unsigned int StartBit = 0,
    unsigned int EndBit = sizeof(Key) * 8
>
auto test_block_radix_sort()
-> typename std::enable_if<Method == 0>::type
{
    using key_type = Key;
    constexpr size_t block_size = BlockSize;
    constexpr size_t items_per_thread = ItemsPerThread;
    constexpr bool descending = Descending;
    constexpr bool to_striped = ToStriped;
    constexpr unsigned int start_bit = (rp::is_unsigned<Key>::value == false) ? 0 : StartBit;
    constexpr unsigned int end_bit = (rp::is_unsigned<Key>::value == false) ? sizeof(Key) * 8 : EndBit;
    constexpr size_t items_per_block = block_size * items_per_thread;

    // Given block size not supported
    if(block_size > test_utils::get_max_block_size())
    {
        return;
    }

    const size_t size = items_per_block * 19;
    const size_t grid_size = size / items_per_block;
    // Generate data
    std::vector<key_type> keys_output;
    if(rp::is_floating_point<key_type>::value)
    {
        keys_output = test_utils::get_random_data<key_type>(size, (key_type)-1000, (key_type)+1000);
    }
    else
    {
        keys_output = test_utils::get_random_data<key_type>(
            size,
            std::numeric_limits<key_type>::min(),
            std::numeric_limits<key_type>::max()
        );
    }

    // Calculate expected results on host
    std::vector<key_type> expected(keys_output);
    for(size_t i = 0; i < size / items_per_block; i++)
    {
        std::stable_sort(
            expected.begin() + (i * items_per_block),
            expected.begin() + ((i + 1) * items_per_block),
            key_comparator<key_type, descending, start_bit, end_bit>()
        );
    }

    // Preparing device
    key_type* device_keys_output;
    HIP_CHECK(hipMalloc(&device_keys_output, keys_output.size() * sizeof(key_type)));

    HIP_CHECK(
        hipMemcpy(
            device_keys_output, keys_output.data(),
            keys_output.size() * sizeof(typename decltype(keys_output)::value_type),
            hipMemcpyHostToDevice
        )
    );

    // Running kernel
    hipLaunchKernelGGL(
        HIP_KERNEL_NAME(sort_key_kernel<block_size, items_per_thread, key_type>),
        dim3(grid_size), dim3(block_size), 0, 0,
        device_keys_output, to_striped, descending, start_bit, end_bit
    );

    // Getting results to host
    HIP_CHECK(
        hipMemcpy(
            keys_output.data(), device_keys_output,
            keys_output.size() * sizeof(typename decltype(keys_output)::value_type),
            hipMemcpyDeviceToHost
        )
    );

    // Verifying results
    ASSERT_NO_FATAL_FAILURE(test_utils::assert_eq(keys_output, expected));

    HIP_CHECK(hipFree(device_keys_output));
}

template<
    class Key,
    class Value,
    unsigned int Method,
    unsigned int BlockSize,
    unsigned int ItemsPerThread,
    bool Descending = false,
    bool ToStriped = false,
    unsigned int StartBit = 0,
    unsigned int EndBit = sizeof(Key) * 8
>
auto test_block_radix_sort()
-> typename std::enable_if<Method == 1>::type
{
    using key_type = Key;
    using value_type = Value;
    constexpr size_t block_size = BlockSize;
    constexpr size_t items_per_thread = ItemsPerThread;
    constexpr bool descending = Descending;
    constexpr bool to_striped = ToStriped;
    constexpr unsigned int start_bit = (rp::is_unsigned<Key>::value == false) ? 0 : StartBit;
    constexpr unsigned int end_bit = (rp::is_unsigned<Key>::value == false) ? sizeof(Key) * 8 : EndBit;
    constexpr size_t items_per_block = block_size * items_per_thread;

    // Given block size not supported
    if(block_size > test_utils::get_max_block_size())
    {
        return;
    }

    const size_t size = items_per_block * 19;
    const size_t grid_size = size / items_per_block;
    // Generate data
    std::vector<key_type> keys_output;
    if(rp::is_floating_point<key_type>::value)
    {
        keys_output = test_utils::get_random_data<key_type>(size, (key_type)-1000, (key_type)+1000);
    }
    else
    {
        keys_output = test_utils::get_random_data<key_type>(
            size,
            std::numeric_limits<key_type>::min(),
            std::numeric_limits<key_type>::max()
        );
    }

    std::vector<value_type> values_output = test_utils::get_random_data<value_type>(size, 0, 100);

    using key_value = std::pair<key_type, value_type>;

    // Calculate expected results on host
    std::vector<key_value> expected(size);
    for(size_t i = 0; i < size; i++)
    {
        expected[i] = key_value(keys_output[i], values_output[i]);
    }

    for(size_t i = 0; i < size / items_per_block; i++)
    {
        std::stable_sort(
            expected.begin() + (i * items_per_block),
            expected.begin() + ((i + 1) * items_per_block),
            key_value_comparator<key_type, value_type, descending, start_bit, end_bit>()
        );
    }

    std::vector<key_type> keys_expected(size);
    std::vector<value_type> values_expected(size);
    for(size_t i = 0; i < size; i++)
    {
        keys_expected[i] = expected[i].first;
        values_expected[i] = expected[i].second;
    }

    key_type* device_keys_output;
    HIP_CHECK(hipMalloc(&device_keys_output, keys_output.size() * sizeof(key_type)));
    value_type* device_values_output;
    HIP_CHECK(hipMalloc(&device_values_output, values_output.size() * sizeof(value_type)));

    HIP_CHECK(
        hipMemcpy(
            device_keys_output, keys_output.data(),
            keys_output.size() * sizeof(typename decltype(keys_output)::value_type),
            hipMemcpyHostToDevice
        )
    );

    HIP_CHECK(
        hipMemcpy(
            device_values_output, values_output.data(),
            values_output.size() * sizeof(typename decltype(values_output)::value_type),
            hipMemcpyHostToDevice
        )
    );

    // Running kernel
    hipLaunchKernelGGL(
        HIP_KERNEL_NAME(sort_key_value_kernel<block_size, items_per_thread, key_type, value_type>),
        dim3(grid_size), dim3(block_size), 0, 0,
        device_keys_output, device_values_output, to_striped, descending, start_bit, end_bit
    );

    // Getting results to host
    HIP_CHECK(
        hipMemcpy(
            keys_output.data(), device_keys_output,
            keys_output.size() * sizeof(typename decltype(keys_output)::value_type),
            hipMemcpyDeviceToHost
        )
    );

    HIP_CHECK(
        hipMemcpy(
            values_output.data(), device_values_output,
            values_output.size() * sizeof(typename decltype(values_output)::value_type),
            hipMemcpyDeviceToHost
        )
    );

    ASSERT_NO_FATAL_FAILURE(test_utils::assert_eq(keys_output, keys_expected));
    ASSERT_NO_FATAL_FAILURE(test_utils::assert_eq(values_output, values_expected));

    HIP_CHECK(hipFree(device_keys_output));
    HIP_CHECK(hipFree(device_values_output));
}

// Static for-loop
template <
    unsigned int First,
    unsigned int Last,
    class T,
    class U,
    int Method,
    unsigned int BlockSize = 256U
>
struct static_for
{
    static constexpr unsigned int end = (end_radix[First] == 0) ? sizeof(T) * 8 : end_radix[First];

    static void run()
    {
        test_block_radix_sort<T, U, Method, BlockSize, items_radix[First], desc_radix[First], striped_radix[First], start_radix[First], end>();
        static_for<First + 1, Last, T, U, Method, BlockSize>::run();
    }
};

template <
    unsigned int N,
    class T,
    class U,
    int Method,
    unsigned int BlockSize
>
struct static_for<N, N, T, U, Method, BlockSize>
{
    static void run()
    {
    }
};

TYPED_TEST(RocprimBlockRadixSort, SortKeys)
{
    using key_type = typename TestFixture::params::input_type;
    using value_type = typename TestFixture::params::output_type;
    constexpr size_t block_size = TestFixture::params::block_size;

    static_for<0, n_sizes, key_type, value_type, 0, block_size>::run();
}

TYPED_TEST(RocprimBlockRadixSort, SortKeysValues)
{
    using key_type = typename TestFixture::params::input_type;
    using value_type = typename TestFixture::params::output_type;
    constexpr size_t block_size = TestFixture::params::block_size;

    static_for<0, n_sizes, key_type, value_type, 1, block_size>::run();
}