will the peformance be worse on DeckLink Cards compared to Graphics Card Output?
I just don‘t know if/how exactly the GPU Acceleration works when I‘m sending the Canvas to a DeckLink output.
The Decklink cards are not GPU and they are not accelerating any calculation. In the contrary, Millumin must transfer the data from the GPU to the RAM then to the Decklink (connected on the motherboard or via a Thunderbolt chassis), so it requires more work.
Hopefully, we optimized this path, and this does not matter that much.
Of course, the advantage of such cards is to be able to output SDI directly, as well as sending sync outputs as for the Decklink 8K Pro.
Thanks for this explanation Philippe.
But that means, that a bigger GPU will still be better because it still does its work before the Data is written to the RAM and than to the DeckLink Cards. Did I understand it right?