how to close threads from the loop

[fiska]

By plaving around with the code, a new issue has arrised,
see the code

#include "MyApp.h"



void MyApp::print()
{
	for (int i = loopStart; i <= loopEnd; ++i)
    {
      if (i % dispFrequency == 0)
      {
          printf( "%d: i = %d\n", threadName, i );
      }
    }
    printf( "%d thread terminating\n", threadName );
}


void MyApp::setName(int &t)
{
#ifdef WITH_SYNCHRONIZATION
    EnterCriticalSection( &m_CriticalSection );  
#endif
	threadName = t;
#ifdef WITH_SYNCHRONIZATION
    LeaveCriticalSection( &m_CriticalSection );   
#endif
}

int main()
{
	

	HANDLE threads[CPU];
	unsigned  uiThread1ID[CPU];


	for(int i = 0; i < CPU; i++)
	{	
		MyApp *o1 = new MyApp(0, 1000000, 20000);
		o1->setName(i); //critical section

		threads[i] = (HANDLE)_beginthreadex( NULL,         // security
									   0,            // stack size
									   MyApp::ThreadStaticEntryPoint,
									   o1,           // arg list
									   0,
									   &uiThread1ID[i] );
		if ( threads[i] == 0 )
			printf("Failed to create thread 1\n");

		//DWORD   dwExitCode;
		//GetExitCodeThread( threads[i], &dwExitCode );  // should be STILL_ACTIVE = 0x00000103 = 259
		//printf( "initial thread %d exit code = %u\n", i, dwExitCode );		
	
	 }

	 WaitForMultipleObjects(CPU, threads, true, INFINITE);

	 printf("This pain has ended!\n");
}

now when I move o1 pointer declaration outside the for loop, exmp in line 37, then there is no interleaving at all, it is executing only the last loop, cor CPU = 5, I get only the last loop i = 4.
But, with the pointer as it is, it works fine, however, creating so many pointers in the loop, I find it inefficient isn't it ?
Furthermore, with the code as it is, when I add the following lines (in the for loop):
delete o1;
o1 = NULL;

My application is behaving strange, and in Win 7 is reporting an error. Why is that? really strange.

***Oh by the way, how does one close HANDLEs in this fashion? In one by one, manual thread creation I used to simply call this
CloseHandle( hth1 );

Or should I still call this in a seperate loop or there is a better way todo so?

thanx