Can the fast DMA move be used implicitely with DirectX Map/Unmap commands?


im currently working with a Blackmagic 8K pro and a Quadro P4000. I used their GPUDirect/DVP example to get data faster to the GPU. Upon testing it versus normal DirectX Map/Unmaps i noticed that the time difference for the whole system was not noticeable. This brought up the question if GPUDirect/DVP has to be used explicitely or if it is used implicitely for DirectX commands as well?