I'm trying to implement a game world similar to this one:
Info: The player is able to dig into the ground.
I already have a plan on how to do the game logic but I didn't come up with a good idea regarding the rendering of such a world. As you can see the whole level consists of thousands of small cubes. Every single one of them can be destroyed at will. For rendering this world my ideas were:
(I haven't tested these so I can only assume they are too slow/bad)
1. Create an Entity for every single cube.
-> Of course not, batch count / memory usage much too high.
2. Create a huge StaticGeometry (SG) with all cubes.
-> Bad, because when a single cube gets destroyed the whole SG has to be rebuilt -> too slow.
3. Create more but smaller SGs by dividing the world in smaller parts, something like an Octree. The SGs have to be small enough so that the player doesn't notice the rebuild after removing a cube.
-> Bad, because the SGs have to be so small that there will be too much of them (high batch count again).
4. Find out every frame which cubes can be seen by the player, then create an Entity for all of those remaining cubes.
-> Still batch count too high?
What can I do to render such a huge amount of small cubes while still being able to remove/create a cube at any place?
Thanks for your attention!
