Parallel programming using Pthreads

Question

I am new in the domain of parallel programming so I decided to fiddle around with the pthread_join() sub routine. I came up with the following code to compute a*X + Y where a is a scalar and X, Y are vectors of some size.

Here is what I have written :

    #include <pthread.h>
    #include <math.h>
    #include <stdio.h>
    #include <stdlib.h>
    #define NUM_THREADS 4
    #define VECTOR_SIZE 65

    struct DAXPYdata
    {
      /* data */
      long a;
      long X[VECTOR_SIZE];
      long Y[VECTOR_SIZE];
    };

    struct DAXPYdata daxpystr;
    void *calcDAXPY(void *);

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

       /*Initialize vectors X and Y an scalar a*/
       daxpystr.a = 57;

    for(vec_index = 0 ; vec_index < 65 ; vec_index++){
        daxpystr.X[vec_index] = vec_index + 1;
        daxpystr.Y[vec_index] = vec_index + 2;
    }

    pthread_t call_thread[NUM_THREADS];
    pthread_attr_t attr;
    int flag;
    long i;
    void *status;

    pthread_attr_init(&attr);
    pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_JOINABLE);

    for(i = 0;i < NUM_THREADS;i++){
        printf("In main() : Creating thread %ld \n", i);
        flag = pthread_create(&(call_thread[i]), &attr, calcDAXPY, (void *)i);
        if(flag == 0){
            printf("ERROR; return code from pthread_create() is %d\n", flag);
            exit(-1);
        }
    }   

    pthread_attr_destroy(&attr);
    for(i = 0; i < NUM_THREADS ; i++) {
       flag = pthread_join(call_thread[i], &status);
       if (flag) {
          printf("ERROR; return code from pthread_join() is %d\n", flag);
          exit(-1);
          }

       printf("main(): Completed join with thread %ld having a status of %ld\n", i, (long)status );
    }
    /* code */
    printf("main(): Program completed, exiting!\n");
    pthread_exit(NULL);
}

void *calcDAXPY(void *thread_id)
{
    int i;
    long tid;

    tid = (long)thread_id;

    printf("Thread %ld starting execution\n", tid);
    for(i = 0 ; i < VECTOR_SIZE ; i++){
        daxpystr.X[i] = daxpystr.a*daxpystr.X[i] + daxpystr.Y[i];
    }

    printf("Result of a*X + Y : ");
    for (i = 0; i < VECTOR_SIZE ; i++)
    {
        printf("%ld ", daxpystr.X[i]);
    }
    pthread_exit((void *)thread_id);
}

The above code throws up an error on pthread_create() subroutine thereby resulting in a single thread execution. Below is the output

In main() : Creating thread 0 
ERROR; return code from pthread_create() is 0
Thread 0 starting execution
Result of a*X + Y : 59 117 175 233 291 349 407 465 523 581 639 697 755 813 871 929 987 1045 1103 1161 1219 1277 1335 1393 1451 1509 1567 1625 1683 1741 1799 1857 1915 1973 2031 2089 2147 2205 2263 2321 2379 2437 2495 2553 2611 2669 2727 2785 2843 2901 2959 3017 3075 3133 3191 3249 3307 3365 3423 3481 3539 3597 3655 3713 3771 logout
Saving session...
...copying shared history...
...saving history...truncating history files...
...completed.

[Process completed]

Any ideas on how to sort this out?

Answer 1

On success, pthread_create() returns 0

so the thread runs, it's just that the error code check should be done like the other way round:

if(flag != 0){
    printf("ERROR; return code from pthread_create() is %d\n", flag);
    exit(-1);
}