Large Mesh at a far distance

So I am trying to do something like KSP, where I have a ground to orbit Level of Detail with the planets. And from theirBuilding a new Universe Video I know they are using scaled space, and local space, where (as i assume) The whole planet is divided in smaller quads. This is all fine but from my understanding They are all still in the same unity scene and rendered( if they’re not invisible( isn’t that automatically decided by the cpu/gpu?)). But now to my question. Even if such a Quad just has 16x16 Vertices, it’s still up thousand if not millions of units away from the scene origin and you and the camera(I know there are two cameras for near and far) Wouldn’t this quad produce a lot of lag and be really inefficient? Or is this not a problem for unity. Generally how far can I put things in unity?

Please also note that I know about the the trick where they move everything back to the origin if the ship passes 6000 units, to reduce the jittering.

• Can you describe why you think a quad a long distance away would cause lag? Or have I misunderstood what you're asking about? – DMGregory Nov 18 '17 at 21:16
• That’s just my question. I mean the quad can be about 1200000 units in size and also about this far away, from the scene origin. Would’nt such a huge object cause lag? – GR00G0 Nov 18 '17 at 21:17
• Geometric size and distance don't themselves create lag in rendering. They can cause imprecision, but an imprecise vertex a long way away is fortunately less noticeable. ;) They can also complicate physics, but at this range you're probably not doing physics simulations on that mesh. Lag coming from rendering is more often due to drawing too many pixels (fill rate) / too many vertices / too many state changes (draw calls, readbacks, etc). So it's unclear to me what kind of performance problem you're trying so solve, or if we might be looking in the wrong place. – DMGregory Nov 18 '17 at 21:29
• No I was just asking for if this would doesnt create performance problems. When I saw that video above i thought i missed something. So I can just have a 600000 radius sphere in real size if i use a proped subdivison algorithm? That just works? – GR00G0 Nov 18 '17 at 21:31
• Well, for certain definitions of "proper" and "just"... ;) Speaking from experience working on (inter)planetary scales on Starlink, making space games isn't easy! But there's no fundamental barrier as long as you're conscious of floating point precision falloff and careful with level of detail. – DMGregory Nov 18 '17 at 21:43