See the question and my original answer on StackOverflow

Here is some sample code to capture the screen with DirectX 9. You shouldn't have to install any SDK (except the standard files that come with Visual Studio, although I didn't tested VS 2010).

Just create a simple Win32 console application, add the following in the stdafx.h file:

  #include <Wincodec.h>             // we use WIC for saving images
  #include <d3d9.h>                 // DirectX 9 header
  #pragma comment(lib, "d3d9.lib")  // link to DirectX 9 library

Here is the sample main implementation

  int _tmain(int argc, _TCHAR* argv[])
    HRESULT hr = Direct3D9TakeScreenshots(D3DADAPTER_DEFAULT, 10);
    return 0;

What this will do is capture 10 times the screen, and save "cap%i.png" images on the disk. It will also display the time taken for this (saving images is not counted in that time, only screen captures). On my (desktop windows 8 - Dell Precision M2800/i7-4810MQ-2.80GHz/Intel HD 4600 which is a pretty crappy machine...) machine, it takes 100 1920x1080 captures in ~4sec, so around 20/25 fps.

  HRESULT Direct3D9TakeScreenshots(UINT adapter, UINT count)
    HRESULT hr = S_OK;
    IDirect3D9 *d3d = nullptr;
    IDirect3DDevice9 *device = nullptr;
    IDirect3DSurface9 *surface = nullptr;
    D3DPRESENT_PARAMETERS parameters = { 0 };
    UINT pitch;
    LPBYTE *shots = nullptr;

    // init D3D and get screen size
    d3d = Direct3DCreate9(D3D_SDK_VERSION);
    HRCHECK(d3d->GetAdapterDisplayMode(adapter, &mode));

    parameters.Windowed = TRUE;
    parameters.BackBufferCount = 1;
    parameters.BackBufferHeight = mode.Height;
    parameters.BackBufferWidth = mode.Width;
    parameters.SwapEffect = D3DSWAPEFFECT_DISCARD;
    parameters.hDeviceWindow = NULL;

    // create device & capture surface
    HRCHECK(d3d->CreateDevice(adapter, D3DDEVTYPE_HAL, NULL, D3DCREATE_SOFTWARE_VERTEXPROCESSING, &parameters, &device));
    HRCHECK(device->CreateOffscreenPlainSurface(mode.Width, mode.Height, D3DFMT_A8R8G8B8, D3DPOOL_SYSTEMMEM, &surface, nullptr));

    // compute the required buffer size
    HRCHECK(surface->LockRect(&rc, NULL, 0));
    pitch = rc.Pitch;

    // allocate screenshots buffers
    shots = new LPBYTE[count];
    for (UINT i = 0; i < count; i++)
      shots[i] = new BYTE[pitch * mode.Height];

    GetSystemTime(&st); // measure the time we spend doing <count> captures
    wprintf(L"%i:%i:%i.%i\n", st.wHour, st.wMinute, st.wSecond, st.wMilliseconds);
    for (UINT i = 0; i < count; i++)
      // get the data
      HRCHECK(device->GetFrontBufferData(0, surface));

      // copy it into our buffers
      HRCHECK(surface->LockRect(&rc, NULL, 0));
      CopyMemory(shots[i], rc.pBits, rc.Pitch * mode.Height);
    wprintf(L"%i:%i:%i.%i\n", st.wHour, st.wMinute, st.wSecond, st.wMilliseconds);

    // save all screenshots
    for (UINT i = 0; i < count; i++)
      WCHAR file[100];
      wsprintf(file, L"cap%i.png", i);
      HRCHECK(SavePixelsToFile32bppPBGRA(mode.Width, mode.Height, pitch, shots[i], file, GUID_ContainerFormatPng));

    if (shots != nullptr)
      for (UINT i = 0; i < count; i++)
        delete shots[i];
      delete[] shots;
    return hr;

Note this code implicitely links to WIC (an imaging library included with Windows for quite a time now) to save the image files (so you don't need the famous D3DXSaveSurfaceToFile that require old DirectX SDKs to be installed):

  HRESULT SavePixelsToFile32bppPBGRA(UINT width, UINT height, UINT stride, LPBYTE pixels, LPWSTR filePath, const GUID &format)
    if (!filePath || !pixels)
      return E_INVALIDARG;

    HRESULT hr = S_OK;
    IWICImagingFactory *factory = nullptr;
    IWICBitmapEncoder *encoder = nullptr;
    IWICBitmapFrameEncode *frame = nullptr;
    IWICStream *stream = nullptr;
    GUID pf = GUID_WICPixelFormat32bppPBGRA;
    BOOL coInit = CoInitialize(nullptr);

    HRCHECK(CoCreateInstance(CLSID_WICImagingFactory, nullptr, CLSCTX_INPROC_SERVER, IID_PPV_ARGS(&factory)));
    HRCHECK(stream->InitializeFromFilename(filePath, GENERIC_WRITE));
    HRCHECK(factory->CreateEncoder(format, nullptr, &encoder));
    HRCHECK(encoder->Initialize(stream, WICBitmapEncoderNoCache));
    HRCHECK(encoder->CreateNewFrame(&frame, nullptr)); // we don't use options here
    HRCHECK(frame->Initialize(nullptr)); // we dont' use any options here
    HRCHECK(frame->SetSize(width, height));
    HRCHECK(frame->WritePixels(height, stride, stride * height, pixels));

    if (coInit) CoUninitialize();
    return hr;

And some macros I used:

  #define WIDEN2(x) L ## x
  #define WIDEN(x) WIDEN2(x)
  #define __WFILE__ WIDEN(__FILE__)
  #define HRCHECK(__expr) {hr=(__expr);if(FAILED(hr)){wprintf(L"FAILURE 0x%08X (%i)\n\tline: %u file: '%s'\n\texpr: '" WIDEN(#__expr) L"'\n",hr, hr, __LINE__,__WFILE__);goto cleanup;}}
  #define RELEASE(__p) {if(__p!=nullptr){__p->Release();__p=nullptr;}}

Note: for Windows 8+ clients, all these (except WIC) should be dropped in favor of the Desktop Duplication API.

Note: for .NET users, there's a C# version here: DirectN