/* ************************************************************************
 * Copyright 2018-2019 Advanced Micro Devices, Inc.
 * ************************************************************************ */

#include "cblas_interface.hpp"
#include "flops.hpp"
#include "norm.hpp"
#include "rocblas.hpp"
#include "rocblas_init.hpp"
#include "rocblas_math.hpp"
#include "rocblas_random.hpp"
#include "rocblas_test.hpp"
#include "rocblas_vector.hpp"
#include "unit.hpp"
#include "utility.hpp"

template <typename T, typename U = T>
void testing_scal_batched(const Arguments& arg)
{
    rocblas_int N           = arg.N;
    rocblas_int incx        = arg.incx;
    U           h_alpha     = arg.get_alpha<U>();
    rocblas_int batch_count = arg.batch_count;

    rocblas_local_handle handle;

    // argument sanity check before allocating invalid memory
    if(N < 0 || incx <= 0 || batch_count <= 0)
    {
        device_vector<T*, 0, T> dx(1);
        if(!dx)
        {
            CHECK_HIP_ERROR(hipErrorOutOfMemory);
            return;
        }

        CHECK_ROCBLAS_ERROR(rocblas_set_pointer_mode(handle, rocblas_pointer_mode_host));
        if(batch_count < 0)
            EXPECT_ROCBLAS_STATUS(
                (rocblas_scal_batched<T, U>)(handle, N, &h_alpha, dx, incx, batch_count),
                rocblas_status_invalid_size);
        else
            CHECK_ROCBLAS_ERROR(
                (rocblas_scal_batched<T, U>)(handle, N, &h_alpha, dx, incx, batch_count));
        return;
    }

    size_t size_x = N * size_t(incx);

    // Naming: dX is in GPU (device) memory. hK is in CPU (host) memory, plz follow this practice

    // Device-arrays of pointers to device memory
    device_vector<T*, 0, T> dx_1(batch_count);
    device_vector<T*, 0, T> dx_2(batch_count);
    device_vector<U>        d_alpha(1);
    if(!dx_1 || !dx_2 || !d_alpha)
    {
        CHECK_HIP_ERROR(hipErrorOutOfMemory);
        return;
    }

    // Host-arrays of pointers to host memory
    host_vector<T> hx_1[batch_count];
    host_vector<T> hx_2[batch_count];
    host_vector<T> hx_gold[batch_count];

    // Host-arrays of pointers to device memory
    // (intermediate arrays used for the transfers)
    device_batch_vector<T> x_1(batch_count, size_x);
    device_batch_vector<T> x_2(batch_count, size_x);

    for(int i = 0; i < batch_count; i++)
    {
        hx_1[i]    = host_vector<T>(size_x);
        hx_2[i]    = host_vector<T>(size_x);
        hx_gold[i] = host_vector<T>(size_x);
    }

    int last = batch_count - 1;
    if((!x_1[last] && size_x) || (!x_2[last] && size_x))
    {
        CHECK_HIP_ERROR(hipErrorOutOfMemory);
        return;
    }

    // Initial Data on CPU
    rocblas_seedrand();
    for(int i = 0; i < batch_count; i++)
    {
        rocblas_init<T>(hx_1[i], 1, N, incx);

        hx_2[i]    = hx_1[i];
        hx_gold[i] = hx_1[i];
    }

    // copy data from CPU to device
    // 1. User intermediate arrays to access device memory from host
    for(int i = 0; i < batch_count; i++)
    {
        CHECK_HIP_ERROR(hipMemcpy(x_1[i], hx_1[i], sizeof(T) * size_x, hipMemcpyHostToDevice));
    }
    // 2. Copy intermediate arrays into device arrays
    CHECK_HIP_ERROR(hipMemcpy(dx_1, x_1, sizeof(T*) * batch_count, hipMemcpyHostToDevice));

    double gpu_time_used, cpu_time_used;
    double rocblas_gflops, cblas_gflops, rocblas_bandwidth;
    double rocblas_error_1 = 0.0;
    double rocblas_error_2 = 0.0;

    if(arg.unit_check || arg.norm_check)
    {
        for(int i = 0; i < batch_count; i++)
        {
            CHECK_HIP_ERROR(hipMemcpy(x_2[i], hx_2[i], sizeof(T) * size_x, hipMemcpyHostToDevice));
        }
        CHECK_HIP_ERROR(hipMemcpy(dx_2, x_2, sizeof(T*) * batch_count, hipMemcpyHostToDevice));
        CHECK_HIP_ERROR(hipMemcpy(d_alpha, &h_alpha, sizeof(U), hipMemcpyHostToDevice));

        // GPU BLAS, rocblas_pointer_mode_host
        CHECK_ROCBLAS_ERROR(rocblas_set_pointer_mode(handle, rocblas_pointer_mode_host));
        CHECK_ROCBLAS_ERROR(
            (rocblas_scal_batched<T, U>(handle, N, &h_alpha, dx_1, incx, batch_count)));

        // GPU BLAS, rocblas_pointer_mode_device
        CHECK_ROCBLAS_ERROR(rocblas_set_pointer_mode(handle, rocblas_pointer_mode_device));
        CHECK_ROCBLAS_ERROR(
            (rocblas_scal_batched<T, U>(handle, N, d_alpha, dx_2, incx, batch_count)));

        // copy output from device to CPU
        for(int i = 0; i < batch_count; i++)
        {
            CHECK_HIP_ERROR(hipMemcpy(hx_1[i], x_1[i], sizeof(T) * size_x, hipMemcpyDeviceToHost));
            CHECK_HIP_ERROR(hipMemcpy(hx_2[i], x_2[i], sizeof(T) * size_x, hipMemcpyDeviceToHost));
        }

        // CPU BLAS
        cpu_time_used = get_time_us();
        for(int i = 0; i < batch_count; i++)
        {
            cblas_scal<T, U>(N, h_alpha, hx_gold[i], incx);
        }
        cpu_time_used = get_time_us() - cpu_time_used;
        cblas_gflops  = batch_count * scal_gflop_count<T>(N) / cpu_time_used * 1e6 * 1;

        if(arg.unit_check)
        {
            unit_check_general<T>(1, N, batch_count, incx, hx_gold, hx_1);
            unit_check_general<T>(1, N, batch_count, incx, hx_gold, hx_2);
        }

        if(arg.norm_check)
        {
            rocblas_error_1 = norm_check_general<T>('F', 1, N, incx, batch_count, hx_gold, hx_1);
            rocblas_error_2 = norm_check_general<T>('F', 1, N, incx, batch_count, hx_gold, hx_2);
        }

    } // end of if unit/norm check

    if(arg.timing)
    {
        int number_cold_calls = 2;
        int number_hot_calls  = 100;
        CHECK_ROCBLAS_ERROR(rocblas_set_pointer_mode(handle, rocblas_pointer_mode_host));

        for(int iter = 0; iter < number_cold_calls; iter++)
        {
            rocblas_scal_batched<T, U>(handle, N, &h_alpha, dx_1, incx, batch_count);
        }

        gpu_time_used = get_time_us(); // in microseconds

        for(int iter = 0; iter < number_hot_calls; iter++)
        {
            rocblas_scal_batched<T, U>(handle, N, &h_alpha, dx_1, incx, batch_count);
        }

        gpu_time_used     = (get_time_us() - gpu_time_used) / number_hot_calls;
        rocblas_gflops    = batch_count * scal_gflop_count<T>(N) / gpu_time_used * 1e6 * 1;
        rocblas_bandwidth = batch_count * (2.0 * N) * sizeof(T) / gpu_time_used / 1e3;

        std::cout << "N,alpha,incx,rocblas-Gflops,rocblas-GB/s,rocblas-us";

        if(arg.norm_check)
            std::cout << ",CPU-Gflops,norm_error_host_ptr,norm_error_device_ptr";

        std::cout << std::endl;

        std::cout << N << "," << h_alpha << "," << incx << "," << rocblas_gflops << ","
                  << rocblas_bandwidth << "," << gpu_time_used;

        if(arg.norm_check)
            std::cout << cblas_gflops << ',' << rocblas_error_1 << ',' << rocblas_error_2;

        std::cout << std::endl;
    }
}

template <typename T, typename U = T>
void testing_scal_batched_bad_arg(const Arguments& arg)
{
    rocblas_int N           = 100;
    rocblas_int incx        = 1;
    U           h_alpha     = U(1.0);
    rocblas_int batch_count = 5;

    rocblas_local_handle handle;

    size_t size_x = N * size_t(incx);

    // allocate memory on device
    device_vector<T*, 0, T> dx(batch_count);
    if(!dx)
    {
        CHECK_HIP_ERROR(hipErrorOutOfMemory);
        return;
    }

    EXPECT_ROCBLAS_STATUS((rocblas_scal_batched<T, U>)(handle, N, nullptr, dx, incx, batch_count),
                          rocblas_status_invalid_pointer);
    EXPECT_ROCBLAS_STATUS(
        (rocblas_scal_batched<T, U>)(handle, N, &h_alpha, nullptr, incx, batch_count),
        rocblas_status_invalid_pointer);
}