@LoopSpace - I agree, I can’t see why you’d need 1000 objects of any size.
I played with your code, and I concede, the shader is faster than a pure mesh approach. ^:)^ I hadn’t expected that, because I’ve often seen advice not to overload shaders with inputs.
@Ignatz The point of my approach is that I’m not continually updating the shader information. So I pass in a fair amount of information but then don’t touch it again. The GPU then does the calculations in parallel, rather than the CPU in series. That leads to the efficiency gain.
@ignatz when i use your discard trick on last loopspace code, then.
1/ the fps is divided by 2.
2/ i still see sprite opaque background parts.
Did i miss stg?
thanks