Cost of thousands of primitive 2d box colliders

[split pivot]

Hello!

I was recently working on a game jam game and I am thinking about taking it further but I have some concerns with cost of performance.

For some context it's a mining game and features hundreds to thousands of 2d box colliders (not featuring rigid bodies) that are generated at runtime to create the level.

I have some concerns about that being really expensive for more complicated levels and game design later down the road.

Some ideas I would have to combat this would be a form of greedy meshing or combining the colliders into one dynamically.

My real question here is this something I should be really worrying myself about as doing a bit of research I see people mentioning it being fine but it just feels wrong to have 45k+ verts on a level for something pretty simple.

Not asking for attention or reviews of my game for the game jam, just here for context incase anyone needed to visualize what I was talking about: https://byteoracle.itch.io/repete-miner

[digital badger]

There's a lot of ways to "solve" this problem but just something worth considering is based off the like 15 seconds I played of your demo you only actually need like 8 colliders right?

If your tiles are being stored in a matrix where you are aware of which tile the player is in you could potentially have a collider for each direction that shifts based on where the player is

and you can use code based logic to determine which tile the collider represents

[split pivot]

Hmm that’s a really smart idea I didn’t think of

[digital badger]

This does have potential downsides eg. things outside of the context of the player such as the bomb or some kind of enemies would need considerations

But in general I think that amount of colliders is mostly fine though

colliders are built to not do too much when nothing is going on with them

but that's a route you could take if you want to do so

nice juice on the tile breaking btw 😄

[split pivot]

Thank you! I was pretty proud of it

[split pivot]

I think I might put together a benchmark basically with my current game jam version and scale up the map various amounts benchmarking performance to see how much I’d really need to worry about it

I’m gonna have to rebuild the game for what I’m wanting to do because game jam codebase 😆

[digital badger]

It really won't matter I don't think. This would purely be something you'd wanna do if you are interested in learning/playing with this kind of problem to solve

[split pivot]

I think a big part of it is the desire of wanting to do something “right”. I have that perfectionist itch in me when I see an issue like this

100% one of my biggest roadblocks though 😆

Anyways thank you for the feedback on it all! It gives me a couple of avenues to check and work on when I’m rebuilding the code in a new project

[digital badger]

Totally relatable and normal feeling!

Just a random strangers two cents but with this one in particular the perfect way would probably end up be some form of calculating it yourself which is a whole mess :P. You'll probably get more gains trying to improve other aspects of the project.