ptheard not working on Ubuntu with matrix - c

Question

this code works correctly on Windows, but does not give results on Ubuntu. What am I supposed to do about it ?
By the way, I use Ubuntu through a virtual machine, but I don't think it has anything to do with it.
#include <pthread.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#define NUM_THREADS 4
#define MATRIX_SIZE 4
int matrixA[MATRIX_SIZE][MATRIX_SIZE];
int matrixB[MATRIX_SIZE][MATRIX_SIZE];
int matrixSerial[MATRIX_SIZE][MATRIX_SIZE]={0};
int matrixParallel[MATRIX_SIZE][MATRIX_SIZE]={0};
//Randomize matrix A and B
void randomizeMatrixAandMatrixB()
{
int i,j;
for(i=0; i<MATRIX_SIZE;i++)
for(j=0; j<MATRIX_SIZE;j++)
{
matrixA[i][j]=rand() % 10;
matrixB[i][j]=rand() % 10;
}
}
//this void is Matrix2D
void printMatrix2D(int mat2D[MATRIX_SIZE][MATRIX_SIZE])
{
int i, j;
for (i = 0; i < MATRIX_SIZE; i++)
{
for (j = 0; j < MATRIX_SIZE; j++)
{
printf("%d ", mat2D[i][j]);
}
printf("\n");
}
printf("\n");
}
//I created this here serailMatrix
void serialMatrixMultiplication()
{
int row, column,i;
for (row = 0; row < MATRIX_SIZE; row++)
for (column = 0; column < MATRIX_SIZE; column++)
for (i = 0; i < MATRIX_SIZE; i++)
matrixSerial[row][column] += matrixA[row][i] * matrixB[i][column];
}
//int step for multiplyMatrix
int step_i = 0;
void *multiplyMatrix(void*arg )
{
int core=step_i++;
int row, column,i;
for (row = core * MATRIX_SIZE / 4; row < (core + 1) * MATRIX_SIZE / 4; row++){
for (column = 0; column < MATRIX_SIZE; column++){
for (i = 0; i < MATRIX_SIZE; i++) {
matrixParallel[row][column] += matrixA[row][i] * matrixB[i][column];
}
}
}
pthread_exit(NULL);
}
//void paraleleling
void parallelMatrixMultiplication()
{
pthread_t *thread;
int i;
for (i = 0; i < NUM_THREADS; i++) {
int* p;
thread=malloc(sizeof(pthread_t));
pthread_create(thread, NULL, multiplyMatrix, (void*)(p));
}
}
int main (int argc, char *argv[])
{
srand(time(0));
//Randomize
randomizeMatrixAandMatrixB();
printf("Matrix A\n");
printMatrix2D(matrixA);
printf("Matrix B\n");
printMatrix2D(matrixB);
//Serial Matrix Multiplication
serialMatrixMultiplication();
printf("Serial Multiplication\n");
printMatrix2D(matrixSerial);
//Parallel Matrix Multiplication
parallelMatrixMultiplication();
printMatrix2D(matrixParallel);
return 0;
}
What should I do to run these codes in the virtual machine? Where is my mistake in the line of code? Since it works fine on Windows, not on linux, it shows the remaining values as 0 from the first line. For example, when we divide it into 4 threads on a 4x4 matrix, the first row looks like this starting from 0. I guess it's not running threads
For Example:
Serial Multiplaction
144 255 1 2
15 18 5 4
125 14 52 46
14 5 1 3
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0

Answer 1

You were lucky it works on windows.
What is missing in you code is synchronisation.
You must wait for threads to finish before displaying the result. You can easily do this with pthread_join
//void paraleleling
void parallelMatrixMultiplication()
{
pthread_t threads[NUM_THREADS];
int params[NUM_THREADS];
int i;
for (i = 0; i < NUM_THREADS; i++) {
params[i] = i;
pthread_create(threads+i, NULL, multiplyMatrix, &params[i]);
}
for (i = 0; i < NUM_THREADS; i++) {
pthread_join(threads[i], NULL);
}
}
As you can see, I also changed the argument given to multiplyMatrix to avoid using step_i: In main thread, an array containing the parameters is prepared.
//int step for multiplyMatrix
void *multiplyMatrix(void*arg )
{
int core=*(int*)arg;
int row, column,i;
for (row = core * MATRIX_SIZE / 4; row < (core + 1) * MATRIX_SIZE / 4; row++){
for (column = 0; column < MATRIX_SIZE; column++){
for (i = 0; i < MATRIX_SIZE; i++) {
matrixParallel[row][column] += matrixA[row][i] * matrixB[i][column];
}
}
}
pthread_exit(NULL);
}
One last thing: if two different threads could modify the same cell in matrixParallel, it would be a good idea to add some mutex.

Answer 2

parallelMatrixMultiplication() kicks off the processing threads then goes straight into printMatrix2D(). You are not waiting for the threads to finish.
You have a memory leak because you are reusing the thread variable in parallelMatrixMultiplication(). You call malloc() but not free().