Byte array +fast blitting

Hi, I''m kind of confused at the moment. The thing is, I have an array of bytes which contains image data, now the thing is I want this data to go to a DirectX surface (an offscreen surface). The problem is there are sooooo many GDI functions to do so I''m a bit confused which will work fastest. What I''m currently doing is: CreateDIBsection() memcpy() (to the DIB I created) BitBlt Is this the fastest way? Are DIB''s faster than DDB''s? And what about palettes (I''m using a monochrome palette I found somewhere at the moment but if at all possible I want to use color). Thanks in advance for anyone who can help me...