First off, sorry for the double post, but I got it working and I love that I was just able to pop it in for the ProceduralMeshComponent.
I have a few questions though.
I am working on a pretty large and complex procedural hex grid terrain.
It gets pretty complex on the highest LOD level.
At the moment I have everyting separated into 8x8 chunks to make use of automated frustum culling
(chunks are actually of variable size, so I can tweak frustum culling vs draw calls)
It actually blows the ProceduralMeshComponent out of the water in terms of performance. But, the more chunks I add the more it slows down though. Which would be expected but frustum culling should also make sure that the meshes out of the view can just be ignored for the most part.
So my main question is: how to get the most out of this for large terrains with high poly count (other than a dynamic LOD system - which I am working on atm). Mesh extraction/polygonizing is quite costly and I’d like to implement a multithreaded solution for it, but not sure how the UpdateFrequency thing works.
Could you give me some hints? 