It may sounds like a stupid question but if I split a model and load all the parts instead of load only one large model can improve performance (memory wise)? I guess not but it’s just a question…
This can in some cases decrease performance by introducing additional draw calls depending on how the materials per mesh are setup.
The all have the same basic material. My idea was to load all of them, merge when they loaded and apply the material to the merged geometry. The main idea is to render in chunks but I started study threejs last month and I dont know to to do this.
No, in general, it’s better to use a single model. But as @titansoftime highlighted, the final performance depends on many factors e.g. the material setup. In any event, it does not make any sense to “render in chunks”. Renderers are actually interested in merging and batching 3D objects so they can produce the final image with as few API commands as possible.