#include <iostream>
#include <numeric>
#include <initializer_list>
#include <cstdlib>
#include <stdlib.h>
#include <half.hpp>

#include "check_err.hpp"
#include "config.hpp"
#include "device.hpp"
#include "host_tensor.hpp"
#include "host_tensor_generator.hpp"
#include "device_tensor.hpp"
#include "device_gemm_xdl_cshuffle.hpp"
#include "element_wise_operation.hpp"
#include "reference_gemm.hpp"
#include "gemm_specialization.hpp"

template <ck::index_t... Is>
using S = ck::Sequence<Is...>;

using PassThrough = ck::tensor_operation::element_wise::PassThrough;

using ADataType        = int8_t;
using BDataType        = int8_t;
using CDataType        = int8_t;
using AccDataType      = int32_t;
using CShuffleDataType = int8_t;

using ALayout = ck::tensor_layout::gemm::RowMajor;
using BLayout = ck::tensor_layout::gemm::ColumnMajor;
using CLayout = ck::tensor_layout::gemm::RowMajor;

static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;

// clang-format off
using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemm_Xdl_CShuffle<
     ALayout,                    // typename ALayout
     BLayout,                    // typename BLayout
     CLayout,                    // typename CLayout
     ADataType,                  // typename ADataType
     BDataType,                  // typename BDataType
     CDataType,                  // typename CDataType
     AccDataType,                // typename GemmAccDataType
     CShuffleDataType,           // typename CShuffleDataType
     PassThrough,                // typename AElementwiseOperation
     PassThrough,                // typename BElementwiseOperation
     PassThrough,                // typename CElementwiseOperation
     GemmDefault,                // GemmSpecialization GemmSpec
     1,                          // index_t NumGemmKPrefetchStage
     256,                        // index_t BlockSize
     256,                        // index_t MPerBlock
     128,                        // index_t NPerBlock
     64,                         // index_t KPerBlock
     16,                         // index_t AK1
     16,                         // index_t BK1
     32,                         // index_t MPerXDL
     32,                         // index_t NPerXDL
     4,                          // index_t MXdlPerWave
     2,                          // index_t NXdlPerWave
     S<4, 64, 1>,                // typename ABlockTransferThreadClusterLengths_AK0_M_AK1
     S<1, 0, 2>,                 // typename ABlockTransferThreadClusterArrangeOrder
     S<1, 0, 2>,                 // typename ABlockTransferSrcAccessOrder
     2,                          // index_t ABlockTransferSrcVectorDim
     16,                         // index_t ABlockTransferSrcScalarPerVector
     16,                         // index_t ABlockTransferDstScalarPerVector_AK1
     1,                          // index_t ABlockLdsExtraM
     S<4, 64, 1>,                // typename BBlockTransferThreadClusterLengths_BK0_N_BK1
     S<1, 0, 2>,                 // typename BBlockTransferThreadClusterArrangeOrder
     S<1, 0, 2>,                 // typename BBlockTransferSrcAccessOrder
     2,                          // index_t BBlockTransferSrcVectorDim
     8,                          // index_t BBlockTransferSrcScalarPerVector
     8,                          // index_t BBlockTransferDstScalarPerVector_BK1
     1,                          // index_t BBlockLdsExtraN
     1,                          // index_t CShuffleMXdlPerWavePerShuffle
     1,                          // index_t CShuffleNXdlPerWavePerShuffle
     S<1, 64, 1, 4>,             // typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
     16>;                        // index_t CShuffleBlockTransferScalarPerVector_NPerBlock
// clang-format on

using ReferenceGemmInstance = ck::tensor_operation::host::
    ReferenceGemm<ADataType, BDataType, CDataType, PassThrough, PassThrough, PassThrough>;

int main(int argc, char* argv[])
{
    bool do_verification = true;
    int init_method      = 1;
    bool time_kernel     = false;

    // GEMM shape
    ck::index_t M = 3840;
    ck::index_t N = 4096;
    ck::index_t K = 4096;

    ck::index_t StrideA = 4096;
    ck::index_t StrideB = 4096;
    ck::index_t StrideC = 4096;

    if(argc == 4)
    {
        do_verification = std::stoi(argv[1]);
        init_method     = std::stoi(argv[2]);
        time_kernel     = std::stoi(argv[3]);
    }
    else if(argc == 10)
    {
        do_verification = std::stoi(argv[1]);
        init_method     = std::stoi(argv[2]);
        time_kernel     = std::stoi(argv[3]);

        M = std::stoi(argv[4]);
        N = std::stoi(argv[5]);
        K = std::stoi(argv[6]);

        StrideA = std::stoi(argv[7]);
        StrideB = std::stoi(argv[8]);
        StrideC = std::stoi(argv[9]);
    }
    else
    {
        printf("arg1: verification (0=no, 1=yes)\n");
        printf("arg2: initialization (0=no init, 1=integer value, 2=decimal value)\n");
        printf("arg3: time kernel (0=n0, 1=yes)\n");
        printf("arg4 to 9: M (256x), N(128x), K(32x), StrideA, StrideB, StrideC\n");
        exit(0);
    }

    auto f_host_tensor_descriptor =
        [](std::size_t row, std::size_t col, std::size_t stride, auto layout) {
            if(std::is_same<decltype(layout), ck::tensor_layout::gemm::RowMajor>::value)
            {
                return HostTensorDescriptor(std::vector<std::size_t>({row, col}),
                                            std::vector<std::size_t>({stride, 1}));
            }
            else
            {
                return HostTensorDescriptor(std::vector<std::size_t>({row, col}),
                                            std::vector<std::size_t>({1, stride}));
            }
        };

    Tensor<ADataType> a_m_k(f_host_tensor_descriptor(M, K, StrideA, ALayout{}));
    Tensor<BDataType> b_k_n(f_host_tensor_descriptor(K, N, StrideB, BLayout{}));
    Tensor<CDataType> c_m_n_host_result(f_host_tensor_descriptor(M, N, StrideC, CLayout{}));
    Tensor<CDataType> c_m_n_device_result(f_host_tensor_descriptor(M, N, StrideC, CLayout{}));

    std::cout << "a_m_k: " << a_m_k.mDesc << std::endl;
    std::cout << "b_k_n: " << b_k_n.mDesc << std::endl;
    std::cout << "c_m_n: " << c_m_n_host_result.mDesc << std::endl;

    switch(init_method)
    {
    case 0: break;
    case 1:
        a_m_k.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5});
        b_k_n.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
        break;
    default:
        a_m_k.GenerateTensorValue(GeneratorTensor_3<ADataType>{0.0, 1.0});
        b_k_n.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5});
    }

    DeviceMem a_m_k_device_buf(sizeof(ADataType) * a_m_k.mDesc.GetElementSpace());
    DeviceMem b_k_n_device_buf(sizeof(BDataType) * b_k_n.mDesc.GetElementSpace());
    DeviceMem c_m_n_device_buf(sizeof(CDataType) * c_m_n_device_result.mDesc.GetElementSpace());

    a_m_k_device_buf.ToDevice(a_m_k.mData.data());
    b_k_n_device_buf.ToDevice(b_k_n.mData.data());

    auto a_element_op = PassThrough{};
    auto b_element_op = PassThrough{};
    auto c_element_op = PassThrough{};

    // do GEMM
    auto gemm     = DeviceGemmInstance{};
    auto invoker  = gemm.MakeInvoker();
    auto argument = gemm.MakeArgument(static_cast<ADataType*>(a_m_k_device_buf.GetDeviceBuffer()),
                                      static_cast<BDataType*>(b_k_n_device_buf.GetDeviceBuffer()),
                                      static_cast<CDataType*>(c_m_n_device_buf.GetDeviceBuffer()),
                                      M,
                                      N,
                                      K,
                                      StrideA,
                                      StrideB,
                                      StrideC,
                                      a_element_op,
                                      b_element_op,
                                      c_element_op);

    if(!gemm.IsSupportedArgument(argument))
    {
        throw std::runtime_error(
            "wrong! device_gemm with the specified compilation parameters does "
            "not support this GEMM problem");
    }

    float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});

    std::size_t flop = std::size_t(2) * M * N * K;
    std::size_t num_btype =
        sizeof(ADataType) * M * K + sizeof(BDataType) * K * N + sizeof(CDataType) * M * N;

    float tflops = static_cast<float>(flop) / 1.E9 / ave_time;

    float gb_per_sec = num_btype / 1.E6 / ave_time;

    std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s, "
              << gemm.GetTypeString() << std::endl;

    c_m_n_device_buf.FromDevice(c_m_n_device_result.mData.data());

    if(do_verification)
    {
        auto ref_gemm    = ReferenceGemmInstance{};
        auto ref_invoker = ref_gemm.MakeInvoker();

        auto ref_argument = ref_gemm.MakeArgument(
            a_m_k, b_k_n, c_m_n_host_result, a_element_op, b_element_op, c_element_op);

        ref_invoker.Run(ref_argument);

        ck::utils::check_err(c_m_n_device_result.mData, c_m_n_host_result.mData);
    }

    return 0;
}