How to use OVERLAPPED to write data at specific offset (win32 programming)

Question

(I'm not native English speaker so sorry for potential mistake :p)

I want to write a Windows tool that can accelerate big files copying just like FastCopy. But get stuck at the very first step: Write data to specific offset of a file via OVERLAPPED (with WriteFile).

Here is my code:

#define _CRT_SECURE_NO_WARNINGS
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <Windows.h>
#include <fileapi.h>

LPOVERLAPPED SetupOvl(DWORD OffsetHigh, DWORD Offset) {
    printf("Ovl??: Got Offset %d\n", Offset);
    LPOVERLAPPED Ovl = (LPOVERLAPPED)malloc(sizeof(OVERLAPPED));
    Ovl->Internal = 0;
    Ovl->InternalHigh = 0;
    Ovl->OffsetHigh = OffsetHigh;
    Ovl->Offset = Offset;
    Ovl->Pointer = NULL;
    Ovl->hEvent = CreateEvent(NULL, TRUE, FALSE, NULL);
    return Ovl;
}

int main()
{
    HANDLE hf,hf2;
    void* buffer2 = (void*)malloc(2 << 16);
    void* buffer = (void*)malloc(2 << 16);
    int readed_number = 0, writed_number = 0;
    DWORD written_number = 0;
    DWORD wn1 = 0, wn2 = 0;

    // OpenThread
    DeleteFileW(L"C:\\Users\\Lev\\Downloads\\dst.txt");
    hf = CreateFileW(
        L"C:\\Users\\Lev\\Downloads\\src.txt",
        GENERIC_READ,
        FILE_SHARE_READ,
        NULL,
        OPEN_EXISTING,
        FILE_ATTRIBUTE_NORMAL,
        0);
    hf2 = CreateFileW(
        L"C:\\Users\\Lev\\Downloads\\dst.txt",
        GENERIC_READ | GENERIC_WRITE,
        0,
        NULL,
        CREATE_NEW,
        FILE_ATTRIBUTE_NORMAL | FILE_FLAG_OVERLAPPED,
        NULL);

    // ReadThread
    BOOL ret = ReadFile(
        hf,
        buffer,
        2 << 15,
        (LPDWORD)&readed_number,
        NULL
    );

    SetFilePointer(hf2, 16, 0, FILE_BEGIN);
    SetEndOfFile(hf2);

    LPOVERLAPPED Ovl1 = SetupOvl(0, 2);
    WriteFile(hf2, buffer, 6, NULL, Ovl1);

    LPOVERLAPPED Ovl2 = SetupOvl(0, 12);
    buffer = (void*)((char*)buffer + (unsigned int)(10));
    WriteFile(hf2, buffer, 4, NULL, Ovl2);

    ret = GetOverlappedResult(hf2, Ovl1, &wn1, TRUE);
    ret = GetOverlappedResult(hf2, Ovl2, &wn2, TRUE);

    printf("Write_A: Write %d Bytes!\n", wn1);
    printf("Write_B: Write %d Bytes!\n", wn2);

    free(Ovl1);
    free(Ovl2);
    CloseHandle(hf);
    CloseHandle(hf2);

    return 0;
}

Output from console:

Ovl??: Got Offset 2
Ovl??: Got Offset 8
Write_A: Write 4 Bytes!
Write_B: Write 5 Bytes!

Content of src.txt: 1234567890abcdefghijklmn

Content of dst.txt: abcd56 (6 charactors and lots of 0x00)

In the code I try to write "123456" at 3rd byte and "abcd" at 13th byte. But it didn't work. I have look up MSDN and some books but found not code examples I need.

By the way, which is the best way to make a multi-thread acceleration? Multi-thread reading or multi-thread writing? Is it possible to use OVERLAPPED I/O to implement multiple threads writing to the same file in parallel? Or does it just write sequentially?

Any helpful suggestion is appreciated!