asdf

asdfsdf​

 

One object will always be better than two because it limits the draw calls. I do make levels made up of pieces like your other option but I take the "level" and combine it in blender and then bring it back to Unity and it treats it as one.

 

In this example It actually shouldn't matter, because number of planes is low, and if they're marked as Static unity will optimize their rendering anyway. Use profiler instead of trying to guess.

https://docs.unity3d.com/Manual/Profiler.html

In this scenario I'd use whatever is more convenient for me, and then alter it only if I ran into performance problems.

 

Second solution should be 25 @ 20x20 planes or 100 @ 10x10 planes. If you densely populate your scene with objects the grid may be a better solution. Really depends on the mechanics you use. Rapidly turning on and off things will burp or choke your frame rate.

 

I say forget the example because it doesn't accurately portray real world scenarios.

There is a lot of factors at play and it comes down to the details of your specific project. I second @neginfinity advice - design for your workflow and only adjust for performance.

If the optimization necessitates a big change in workflow that would mean a major do-over, then setup a proper test that is as close to the real thing as possible.

An example: in my current project, we have big open world. About 100 square kilometers. Conventional wisdom says to using streaming techniques for terrain tiles. So we did this. Turns out, considering all the unique factors of our project, it is actually significantly more performant to not stream the terrains - instead load them all at once. If we had tested the simple way first, would have saved a lot of time and money.

 

I am like one large plan but it depend on plan.

 

Not necessary if you mark your objects as static, Unity will combine them anyway to reduce draw calls.

 

Since static batching still does culling for each mesh, it’s not quite the same as it being one giant mesh though.