#include<iostream>
#include<vector>
#include<cmath>
#include<fstream>
#include<omp.h>
#include<chrono>         

using namespace std::chrono;
int main()  
{
  const int n=100000000; 
  std::ofstream fileout("timing");
  double time,reftime=0.0;
  const int ncores = 12;
  std::cout<<ncores<<"\n";
  for(int nc=1; nc!=ncores+1; nc++) { 
    omp_set_num_threads(nc); // override OMP_NUM_THREADS 
    double sum=0;
    auto t0 = high_resolution_clock::now();
#pragma omp parallel for reduction(+:sum)
    for (int i = 0; i < n; i++) sum+=1/cos(i);//thread sums put together in the end
    auto t1 = high_resolution_clock::now();		        
    auto d = duration_cast<milliseconds>(t1-t0);
    time = d.count()/1000.0; 
    std::cout<<"sum = "<<sum<<"   cores "<<nc<<"  timing "<<time<<" s"<<"\n";
    if(nc==1) reftime = time;
    fileout<<nc<<" "<<time<<" "<<reftime/time<<"\n";      
  }
  fileout.close();
}