#include <omp.h>
#include <iostream>
#include <ctime>
#include <locale>
#include <cstdlib>
using namespace std;

double normOfMtx(double** C, int Dimension)
{
    double sum = 0;
    for (int i = 0; i < Dimension; i++)
        for (int j = 0; j < Dimension; j++)
            sum += C[i][j] * C[i][j];
    return sqrt(sum);
}

void multiplyVer1(double **A, double **B, int Dimension, int numThreads, long q)
{
    int i, j, k;
    double** C = new double*[Dimension];
    for (i = 0; i < Dimension; i++)
    {
        C[i] = new double[Dimension];
        for (j = 0; j < Dimension; j++)
            C[i][j] = 0;
    }
    double s = 0;
    int t1 = clock();

for (int l = 0; l < q; l++)
{
#pragma omp parallel for num_threads(numThreads)
        for (i = 0; i < Dimension; i++)
        {
            #pragma omp parallel for num_threads(numThreads)
            for (j = 0; j < Dimension; j++)
            {
                for (k = 0; k < Dimension; k++)
                {
                    /*s += A[i][k] * B[k][j];*/
                    C[i][j] += A[i][k] * B[k][j];
                }
                /*C[i][j] = s;
                s = 0;*/
            }
        }
}
    int t2 = clock();
    /*for (i = 0; i < Dimension; i++)
    {
        for (j = 0; j < Dimension; j++)
            cout << C[i][j] << " ";
        cout << endl;
    }*/
    cout << "Total " << ((float)(t2 - t1)) / CLOCKS_PER_SEC << " sec." << endl;
    cout << "Norm is " << normOfMtx(C, Dimension) << endl;
    system("pause");
    for (int q = 0; q < Dimension; q++)
        delete[] C[q];
}

void multiplyVer2(double **A, double **B, int Dimension, int numThreads, long q)
{
    int i, j, k;
    double** C = new double*[Dimension];
    for (i = 0; i < Dimension; i++)
    {
        C[i] = new double[Dimension];
        for (j = 0; j < Dimension; j++)
            C[i][j] = 0;
    }
    double s;
    int t1 = clock();
for (int l = 0; l < q; l++)
{
    for (i = 0; i < Dimension; i++)
    {
#pragma omp parallel for /*shared(A, B, C)*private(j, k)*/ num_threads(numThreads)
        for (j = 0; j < Dimension; j++)
        {
            for (k = 0; k < Dimension; k++)
            {
                C[i][j] += A[i][k] * B[k][j];
            }
        }
    }
}
    int t2 = clock();
    /*for (i = 0; i < Dimension; i++)
    {
        for (j = 0; j < Dimension; j++)
            cout << C[i][j] << " ";
        cout << endl;
    }*/
    cout << "Total " << ((float)(t2 - t1)) / CLOCKS_PER_SEC << " sec." << endl;
    cout << "Norm is " << normOfMtx(C, Dimension) << endl;
    system("pause");
    for (int q = 0; q < Dimension; q++)
        delete[] C[q];
}

void multiplyVer3(double **A, double **B, int Dimension, int numThreads, long q)
{
    int i, j, k;
    double** C = new double*[Dimension];
    for (i = 0; i < Dimension; i++)
    {
        C[i] = new double[Dimension];
        for (j = 0; j < Dimension; j++)
            C[i][j] = 0;
    }
    double s;
    int t1 = clock();
for (int l = 0; l < q; l++)
{
#pragma omp parallel for /*shared(A, B, C) private(j, k)*/ num_threads(numThreads)
    for (i = 0; i < Dimension; i++)
    {
        for (j = 0; j < Dimension; j++)
        {
            for (k = 0; k < Dimension; k++)
            {
                C[i][j] += A[i][k] * B[k][j];
            }
        }
    }
}
    int t2 = clock();
    /*for (i = 0; i < Dimension; i++)
    {
        for (j = 0; j < Dimension; j++)
            cout << C[i][j] << " ";
        cout << endl;
    }*/
    cout << "Total " << ((float)(t2 - t1)) / CLOCKS_PER_SEC << " sec." << endl;
    cout << "Norm is " << normOfMtx(C, Dimension) << endl;
    system("pause");
    for (int q = 0; q < Dimension; q++)
        delete[] C[q];
}

int main()
{
    setlocale(LC_ALL, "Russian");
    cout << "Перемножение матриц" << endl << "Введите размерность квадратной матрицы" << endl;
    int Dimension;
    cin >> Dimension;
    cout << "Введите количество повторов для умножения" << endl;
    long q;
    cin >> q;
    srand(time(nullptr));

    long long i, j, k, l;
    double** A = new double*[Dimension];
    double** B = new double*[Dimension];
    for (i = 0; i < Dimension; i++)
    {
        A[i] = new double[Dimension];
        B[i] = new double[Dimension];
    }

#pragma omp parallel for shared(A, B) private(i, j) num_threads(4)
    for (i = 0; i < Dimension; i++)
    {
        //cout << "[";
        for (j = 0; j < Dimension; j++)
        {
            A[i][j] = (rand() % 10 - 5) / 10.0;
            //cout << A[i][j] << " ";
        }
        //cout << "][";
        for (j = 0; j < Dimension; j++)
        {
            B[i][j] = (rand() % 10 - 5) / 10.0;
            //cout << " " << B[i][j];
        }
        //cout << "]" << endl;
    }

    //------------------------------------------—
    multiplyVer1(A, B, Dimension, 4, q);
    multiplyVer2(A, B, Dimension, 4, q);
    multiplyVer3(A, B, Dimension, 4, q);
    //------------------------------------------—
    for (int i = 0; i < Dimension; i++)
    {
        delete A[i];
        delete B[i];
    }
    return 0;
}