Performance? - material with 6 material functions VS model divided into 6 parts, one material each.

Development Rendering
, , , ,
1

Hi guys,

I have a conundrum, and I don’t know what would be more efficient. Currently we have layered materials with some up to or around 6 mf’s, most of them relatively simple base materials with options for tiling, normal strength and a detail normal.

Would splitting the model up into multiple uv sets and applying a base material per uv be worse off or better?

Would using multiple UV-sets on 1 model with one material and use the texcoord node to point the UV sets to material functions be more efficient than splitting up a model into different chunks? Where every chunk uses 1 material (and thus fewer material functions would be required per material).

So basically a more complex material versus more model parts and thus drawcalls?

Best regards, Käy Vriend.

2

So basically a more complex material
versus more model parts and thus
drawcalls?

Essentially it is situational and there is no all-in-one response. You should do few profiling attempts to decide.

3

Well, thanks. I did try and sketch a specific scenario in the description, hoping for a more specific reply though. Curious to see what the Unreal Engine devs have to say.