I'm trying out the Direct2D RenderTarget property of C++ Builder's TDirect2DCanvas to benchmark bitmap manipulation using the GPU. This is most impressive compared to GDI. However, I cannot figure out how to get the result back into CPU memory so I can use it for further CPU processing. Did read about using WIC but apparently you loose most of the hardware acceleration. Windows 8 seem to offer a new bitmap with map functionality, but I want to use (standard) Windows 7. Also C++ Builder does not offer the header files for these new functions.
Given a TDirect2DCanvas, or so you wish a Direct2D RenderTarget or a ID2D1Bitmap, how do I copy the bitmap back to CPU memory?