- Home /
Many draw calls w\ low verts VS little draw calls w\ lots of verts
For starters I'm targeting Mobile, specifically Android and IOS.
I'm going around improving performance and it occurred to me that the majority of my static objects (like rocks/background objects/etc) I can combine into one object to reduce draw calls. But my concern is - is it more efficient to have one HUGE object (so 1 draw call) that contains a lot of tris (lets say for an example 20k) or is it better to have maybe 10 draw calls with each draw call having 2k tris?
Either way I'm going to have the same amount of tris, but is it faster for a GPU to do small tri objects or just go for lowest draw calls possible?
Answer by Crystalline · Oct 27, 2013 at 10:12 PM
Having a big (many tris) object makes it easier for the GPU to render it because ,yes it does that in one drawcall, however make sure you only join objects that are pretty colse one to each other. Joining too many objects from all over the map will make the frustum culling inefficient as the GPU will be rendering the whole mesh/object each frame.
So, its best to get to a balance , join close objects..
Ah thanks. Didn't know about the frustum culling so thanks for that information.