Pthread Matrix Multiplication (broken)

#include <pthread.h>
#include <stdio.h>
#include <stdlib.h>

int **x,**y, **z;
int NUM_THREADS;
int SIZE;

void error()
{
    fprintf(stderr,"Malloc fail!");
    exit(3);
}

void generateMatrix()
{
    int i,j;
    if((x = malloc(sizeof(int*)*SIZE))==NULL)error();
    if((y = malloc(sizeof(int*)*SIZE))==NULL)error();
    if((z = malloc(sizeof(int*)*SIZE))==NULL)error();

    for(i=0;i<SIZE;i++)
    {
        if((x[i] = malloc(sizeof(int)*SIZE))==NULL)error();
        if((y[i] = malloc(sizeof(int)*SIZE))==NULL)error();
        if((z[i] = malloc(sizeof(int)*SIZE))==NULL)error();
    }

    for(i=0;i<SIZE;i++)
        for(j=0;j<SIZE;j++)
        {
            x[i][j] = 1;
            y[i][j] = 1;
        }
}

void *compute(void *arg)
{
    int tid = (int) arg;
    int i,j,k,BOUND;

    /* Calculate boundary */
    BOUND = (SIZE/NUM_THREADS)*(tid + 1);

    printf("tid:%d\tbound:%d\n",tid,BOUND);

    for(i=0;i<SIZE;i++)
    {
        j=tid*(SIZE/NUM_THREADS);
        for(;j<BOUND;j++)
        {
            z[i][j] = 0;
            for(k=0;k<SIZE;k++)
            {
                z[i][j] += x[i][k] * y[k][j];
                if(z[i][j] == -7)
                    printf("z[%d][%d]=%d,x[%d][%d]=%d,y[%d][%d]=%d\n",i,j,z[i][j],i,k,x[i][k],k,j,y[k][j]);
            }
        }
    }

    pthread_exit((void *)arg);
}

int main(int argc, const char *argv[])
{
    int i,j, ret;
    FILE *ofp;
    pthread_t threads[NUM_THREADS];
    pthread_attr_t attr;
    void *status;

    if(argc == 3)
    {
        SIZE = atoi(argv[1]);
        NUM_THREADS = atoi(argv[2]);
    }
    else
        exit(-2);

    generateMatrix();

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

    for(i=0;i<NUM_THREADS;i++)
    {
        printf("Thread %d is created\n",i);
        ret = pthread_create(&threads[i], NULL, compute, (void *) i);
        if(ret)
        {
            fprintf(stderr, "Thread creation failed!\n");
            exit(ret);
        }
    }

    //pthread_attr_destroy(&attr);
    for(i=0;i<NUM_THREADS;i++)
    {
       ret = pthread_join(threads[i], &status);
       if(ret)
       {
           fprintf(stderr,"Thread join failed! status:%d\n",ret);
           exit(ret);
       }
       printf("Thread %d join with main thread with status=%ld\n",i,(long)status);
    }

    ofp = fopen("output.txt","w");
    if(ofp == NULL)
    {
        fprintf(stderr, "Can't open output file");
        exit(1);
    }

    for(i=0;i<SIZE;i++)
    {
        for(j=0;j<SIZE;j++)
            fprintf(ofp,"%d ", z[i][j]);
        fprintf(ofp,"\n");
    }

    fclose(ofp);
    for(i=0;i<SIZE;i++)
    {
        free(x[i]);
        free(y[i]);
        free(z[i]);
    }
    pthread_exit(NULL);
    return 0;
}