#include "bsp_3_lib_bench_par.h"
#include <cmath>
#include <iostream>

using namespace std;

int main()
{
    benchmark_B(8000,8000,400);
    benchmark_B_par(8000,8000,400);
    benchmark_C(4000,4000,4000,1);
    benchmark_C_par(4000,4000,4000,1);
    benchmark_D(1e4,1e5,15);
    benchmark_D_par(1e4,1e5,15);
    
    // comparing the time for sum and inner product with and without parallelization
    cout << "\nComparing the runtime (in sec) for inner product and sum with and without parallelization\n\n";
    for(int k=3; k<=8; ++k)
    {
		
		int N = pow(10,k);
		cout << "k = " << k << "	N = " << N << endl;
		vector<double> v1(N,1.0/N);
		vector<double> v2(N,N);
		
		auto tstart = omp_get_wtime();
		double s = scalar(v1, v2);		
		auto t_diff_scalar = omp_get_wtime() - tstart;
		
		tstart = omp_get_wtime();
		double sp = scalar_par(v1, v2);		
		auto t_diff_scalar_par = omp_get_wtime() - tstart;
		
		tstart = omp_get_wtime();
		double su = sum(v1);		
		auto t_diff_sum = omp_get_wtime() - tstart;
		
		tstart = omp_get_wtime();
		double sup = sum_par(v1);		
		auto t_diff_sum_par = omp_get_wtime() - tstart;
		
		cout << "sum	" << t_diff_sum << "	inner_prod	" << t_diff_scalar << endl;
		cout << "sum_par	" << t_diff_sum_par << "	inner_prod_par	" << t_diff_scalar_par << endl << endl;;
	}

    return 0;
}