#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/time.h>
#define ARRAY_SIZE 1024
#define NUMBER_OF_TRIALS 1000000

double dtime() {
    double tseconds = 0.0;
    struct timeval my_t;
    gettimeofday(&my_t, NULL);
    tseconds = (double)(my_t.tv_sec + my_t.tv_usec * 1.0e-6);
    return (tseconds);
}

int get_model_name(char *mname) {
    FILE *fp;
    fp = fopen("/proc/cpuinfo", "r");
    if (fp == NULL) {
        strcpy(mname, "(/proc/cpuinfo is not readable)\n");
        return(1);
    }
    /* model name should be on the fifth line */
    for (int i=0; i < 5; i++) fgets(mname, 80, fp);
    fclose(fp);
    return(0);
}

int main(int argc, char *argv[]) {

    /* Declare arrays small enough to stay in L1 cache.
       Assume the compiler aligns them correctly. */
    double a[ARRAY_SIZE], b[ARRAY_SIZE], c[ARRAY_SIZE];
    int i, t, rc;
    double m = 1.5, w1, w2, d = 0.0;
    char modelname[80];

    /* Initialize a, b and c arrays */
    for (i=0; i < ARRAY_SIZE; i++) {
        a[i] = 0.0; b[i] = i*1.0e-9; c[i] = i*0.5e-9;
    }

    /* Perform operations with arrays many, many times */
    w1 = dtime();
    for (t=0; t < NUMBER_OF_TRIALS; t++) {
        for (i=0; i < ARRAY_SIZE; i++) {
            a[i] += m*(m*b[i] + c[i]);
        }
    }
    w2 = dtime();

    /* Print total time and processor type used in the run.
       Print a result so array ops aren't optimized away. */
    for (i=0; i < ARRAY_SIZE; i++) d += a[i];
    printf("d = %f   time = %f\n", d, w2 - w1);
    rc = get_model_name(modelname);
    if (rc == 0) printf("%s", modelname);
}