How much does an invisible mesh affect performance?

Let’s say I’ve half a million tiny mesh trees spread across a 16000x16000 terrain map. They all have CanQuery set to false and are invisible. A localscript would check players surroundings every now and then and make nearby trees visible. Would the virtually inexistent hundreds of thousands of trees make a significant decrease to performance on ping or fps?

This is a hypothetical question, I don’t have half a million trees anywhere. Yet. :lying_face:

1 Like

bros just waiting for the okay :rofl:,

honeslty It probably might but i guess it depends on the hardware of your players, if its maybe more mobile players then it may be a bad idea

I think having half a million instances will affect performance, even if those instances don’t affect rendering or physics/querying.

Instead you could store only the positions (also maybe rotation) of those half a million trees in a table, then continuously move 100 trees to the positions close to the player.

Yeah I had this in mind since CanQuery would make them unreadable anyway. :thinking: But I’m curious if anyone has any statistics before I try that out in practice lol

Giant Tables will affect your performance Extremely. Instead of that, create a sphere that permanently stays where the player is(connected via RunService Renderstepped) and then make all the trees touching that sphere visible and everything that leaves this sphere invisible. Done

And why don’t you use Streaming lol?

That would also stream out the terrain, wouldn’t it? :thinking:

It would, but you don’t notice.

it depends on the hardware… It might lag.

Of course you do if its not isometric

To be sure, it’s easier to do a benchmark yourself.
Duplicate thousands of objects and test performance in a loop.
Here’s an idea of how to benchmark: