I heard custom node can be intensive and should be used only when you can’t replicate the same functionality with premade nodes, but I was wondering if this is the case…(I looked around and I couldn’t find any logic gate node, is it normal and you need to build yourself or it’s my fault?).
Generally this is true, but there are exceptions. For example, dynamic branching can only be done within the custom node, and in some cases could result in a more optimized material.
I suspect measuring any difference between these two graphs would be challenging. I would be curious as to what the result is, but I doubt it makes any practical difference.
Yes the material graph does some optimizations, but in my opinion the main reason to use custom nodes minimally is that aside from HLSL intrinsics, everything else is subject to change so there’s no guarantee your code will work when you change versions, a lesson I’ve had to learn repeatedly.
1 Like
I asked because I was planning to use hundreds of this material instance for a “Billboard” effect and didn’t wanted it to be to much intensive on the computer to render (there are probably many more changes I can do to the material to make it faster, but this OR gate especially makes the node graph looks really ugly and difficult to understand)
Test it