#define _GNU_SOURCE
#include <sys/types.h>
#include <omp.h>
#include <stdio.h>
#include <math.h>
#include <time.h>
#include <sys/time.h>
#include <unistd.h>  /* sleep() */
#include <sched.h>
#include <asm/unistd.h>
#include <sys/syscall.h>
#define CHUNKSIZE 128
#define N     128000
#define Ninner     2000

#define gettid() syscall(SYS_gettid)

#define ALIGN(n) __attribute__((aligned(n)))
float a[N] ALIGN(128), b[N] ALIGN(128), c[N] ALIGN(128);

void process(int threads) {
    int chunk;

    clock_t c0, c1;
    struct timeval t0, t1;
    

    /* Some initializations */
    for (int i=0; i < N; i++)
      a[i] = b[i] = i * 1.0;
    chunk = CHUNKSIZE;
    // printf("running with %i threads: ",threads);


    c0 = clock();
    gettimeofday(&t0,NULL);

    #pragma omp parallel num_threads(threads) shared(a,b,c,chunk)
    {
      int i, j;
      #pragma omp for schedule(dynamic,chunk) nowait
      for (i=0; i < N; i++) {
        for (j=0; j<Ninner; j++) {
            c[i] = sqrt(sqrt(a[i])) + sin(cos(b[i])) + sqrt(j);
        }
      }

    }  /* end of parallel section, implicit barrier here(?) */

    c1 = clock();
    gettimeofday(&t1,NULL);

    float sum=0;
    for(int i=0; i<N; i++) sum += c[i];

    //printf("difftime=%.4lf s\n",difftime(end, start));
    double runtime = (double)(t1.tv_sec-t0.tv_sec) + (double)(t1.tv_usec-t0.tv_usec)/1e6;

    printf("%d %f %f %f\n", threads, runtime, (c1-c0)*1./CLOCKS_PER_SEC, sum);
    return;
}

int main ()  
{
    int i;
    for (i=1;i<=sysconf(_SC_NPROCESSORS_ONLN);i<<=1) {
        process(i);  
    }
    return 0;
}