Basic Matrix multiplication with pthreads c++ error - c++

Question

I have a matrix multiplication code that I am supposed to process in parallel. I have a code here that I believe should work but does not. It either causes segmentation faults or gives me all gibberish values. Can any one help? Thanks in advance.
//*******************STRUCTS AND GLOBAL VARIABLES*****************************//
struct Matrix
{
int d[SIZE][SIZE];
};
Matrix* matrix_addr[SIZE]; // array to store the address of the matrices
int n;
int m;
pthread_mutex_t my_mutex = PTHREAD_MUTEX_INITIALIZER;
pthread_cond_t my_cond = PTHREAD_COND_INITIALIZER;
//****************************THREAD******************************************//
void* calcTerm(void* arg)
{
pthread_mutex_lock(&my_mutex);
int sum = 0;
Matrix* m0 = (Matrix*) arg;
Matrix* m1 = (Matrix*) ((int*)arg + 1);
Matrix* m2 = (Matrix*) ((int*)arg + 2);
cout << endl << "Print\n" << endl;
print (m0);
for (int i = 0; i < SIZE; ++i)
{
cout << "\ni = " << i << "\tn = " << m1->d[n][i] << "\tm = " << m2->d[i][m] << endl;
sum = sum + (m1->d[n][i] * m2->d[i][m]);
}
cout << endl << endl << sum << endl;
m0->d[n][m] = sum;
pthread_mutex_unlock(&my_mutex);
cout << endl << "Going out of thread\n" ;
pthread_exit(NULL);
}
//********************************MAIN****************************************//
int main()
{
Matrix m0, m1, m2; //Matrices are 3x3;
// m0 <= m1 * m2
pthread_t id[9]; // 3x3 matrix multiplication requires 9 threads.
matrix_addr[0] = &m0; // the pointers to the matrices are stored here.
matrix_addr[1] = &m1;
matrix_addr[2] = &m2;
n = m = 0; // initialize the global variable
srand(time(NULL)); // seed rand()
for (int i = 0; i < SIZE; i++)
{
for (int j = 0; j < SIZE; j++)
{
m0.d[i][j] = 0; // m0 is being cleared for the output
m1.d[i][j] = rand()%10; // m1 and m2 are generated with rand()
m2.d[i][j] = rand()%10;
}
}
//display the input matrices
cout << "MATRIX 1:\n\n";
print (&m1);
cout << "\nMATRIX 2:\n\n";
print (&m2);
cout << "\nMATRIX 3:\n\n";
print (&m0);
for (int i = 0; i < SIZE*SIZE; i++) // run all the threads for calculating each output
{
m = i % SIZE;
n = i / SIZE;
cout << endl << "Going in to thread " << i << " with n = " << n << " and m = " << m;
pthread_create(&id[i], NULL, calcTerm, (void*) matrix_addr);
cout << endl << "Out of thread " << i ;
//pthread_join(id[i], NULL);
}
//pthread_cond_wait(&my_cond, &my_mutex);
cout << endl << endl;
print_result (&m0, &m1, &m2);
return 0;
}
It seems like the calcTerm thread does not take the correct pointers or something. It calculates gibberish values, but the final output at the end of main prints the same matrices I started off with.
Thanks again.