Hi,
I have a performance/optimization question, it looks like most composite operators take 0.05 to 0.08 ms to cook, and that’s the same for small or big input pictures.
In my current project I had a naive approach where I clone containers and do lots of composite operations on small pictures and then merge them into a bigger picture, but I guess I should assemble things in bigger pictures first and do the operations then to optimize speed, if I’m cpu limited.
Did I get this right?
Though I guess if the pictures are really huge, then I’ll hit a gpu bottleneck (which I had in another project)
Thanks
Vincent!