Hello guys, thanks for the answers. Sorry for the ambiguity.
What I am trying to do is multiplying sprites, similar to bacterial patterns. They sort of interact with the environment, which could be both the map and other sprites, and should spawn a child at a certain event. I guess you can imagine that the number of sprites increases quickly.
I see the limitations by trying this all on the existing screen bounds, I'd rather have a larger map where I could scroll around, but ultimately, zoom to a level where I could have an overview of the entire world. That way I could have way more sprites than fit on one screen, but once I zoom out, it would be necessary to view all of them.
I built a quick prototype in order to check a little for myself, and hit serious performance issues at around 5000 sprites - talking about 20 seconds calculation time on each iteration through the sprite array and rapidly increasing as the total number of sprites still increases.
This prototype is in no way optimized and surely I haven't used advanced practices, as I am completely new to Cocos2D, but I am unsure if this is even the right tool for what I have in mind.
It's certainly amazing how much functionality is built in and how quickly I was able to code the AI, but maybe an idea like mine just doesn't fit into the framework. I am not even sure that I would call the project a game in the true sense :)
So do you guys think I should use cocos2D for that?
If so, do you have recommendations for best practices I should read about?
For example, the hint that overlapping sprites cost a lot of processing time is very valuable, although I haven't planned on overlapping, but it's good to know in advance then find out the painful way.
Should I use Box2D for collision detection? In the prototype, I created a multidimensional array which stores positions for a quick lookup...
