Can we discuss the performance of levels vs components per landscape?
I have a 16384x8192 res digital elevation model, and am wondering about it will perform in different senarios, and what the hit for level streaming is vs levels with a larger number of components.
For instance, with this model, I could have two 8192x8192 vert levels, or I could further subdivide that so that I could have a number of say, 16 1008x1008 vert levels.
I’m guessing that having less components per level would help performance at smaller scales (what would generally be reasonable for a game engine), as there are less draw calls, but mean that there are more places where level loading has to happen… (unless I’ve missunderstood how level streaming works), but if we wanted to view the whole landscape (which we do for my usecase), our performance would get worse as we got further in?
Does this sound right?