DOTS Advice

[primal torrent]

Being a DOTS novice, i thought about trying a hybrid approach, where I have a parent with two children, one is the Renderer only and the other is for components such as colliders.

My plan was to grab all those Renderers, MeshCombine them, then stick the resulting mesh in a subscene to take advantage of the Entities system for visuals whilst keeping the collider components seperate to allow me to keep using the standard physics system.

My question is, would it even be worth doing this? does having the renderers as entities give any performance benefit whatsoever? especially with very large meshes with high vertices counts?

This is of course separate to the performance benefits inherent to Jobs and Burst, was trying to perhaps find a method of leveraging ECS without needing to dive into actual ECS code (yet).

Curious to hear peoples thoughts on this.
Thanks

[rustic violet]

does having the renderers as entities give any performance benefit whatsoever? especially with very large meshes with high vertices counts?

I think the power of ECS is in numbers; large meshes with high vertex counts will be processed engine side anyway, I doubt you'd gain a lot from having that in ECS. If you merge them, a few gameobjects with large meshes won't make a big difference

[bright jasper]

Usually people do it vice versa - ECS gameplay and GameObject renderers (for animations)