About The New Hardware Renderer

Development
1

I’ve briefly thought about the new hardware renderer still planned for Tiled after reading about it via GitHub, and I find it really neat about how it’ll improve performance and make it far less sloggish for anyone wanting to go make super crazy-huge maps for games.

I calculated the amount of time it’ll take to be implemented in my mind after reading it, which is 350 hours from what I’ve read, and it narrows down to about 15 days, which is 2 weeks and almost a day, 10 hours above 350, to be accurate.

Figuring out the amount of days to work on it was easy for me, but I already know the actual work isn’t.

I don’t know the present progress status of it, but I can only guess some or a bit of it is made… maybe?

I already know work’s already focused on 1.12 right now, but I may as well ask you this, anyway; would focus on the new renderer be part of 1.12, or for a later update after?

Thought I would ask this for discussion, and in case it seems to be forgotten again.

Well, whatever the matter, I’m very hopeful it’ll be done. I’m still looking forward to it. Thanks again!

2

Is this AI generated? The math is off quite a bit. We’re looking at a project of at least 2 months. That’s assuming the estimate is accurate, but software development time is usually underestimated and my estimations are no exception to that. :slight_smile:

The only progress that has been made is part of “Tiled Quick”, which is part of the repository in src/tiledquick, src/libtiledquick and src/tiledquickplugin. At the moment it’s just a map viewer.

Yes, my focus for 1.12 are the new Properties view and support for array properties, but anybody is free to work on other tasks. I don’t think a new hardware renderer can still make it into 1.12 though, but there’s always the next release.

Are you interested in working on this feature?

3

Nope. 100% pure Human. I only used my real brain to figure out the estimate. But, yeah, I try to be closely accurate as possible. I counted by how many hours there are in a day, and I kinda thought I was accurate, but, barring any weekend days and nighttime hours, if needed, then… yeah.

OK… Nice of you to share. Don’t know what Tiled Quick is, anyway. I’m guessing, maybe a new feature or what?

Well, it was worth asking, anyway.

I’m only a newcomer to the forums. No familiarity nor experience about that outside of here.

4

I quickly forgot to ask one, small question… about the effects of the new renderer.

Once the new renderer is implemented, would it also change the way it handles fixed maps’ sizes?

I’m not counting those made with infinite maps in that matter.

My reason for this question… I briefly, and curiously, thought the caps on the fixed maps’ sizes were based on existing installed RAM amounts in computers(honestly, I didn’t actually think of asking that before, but, presently, do they, anyway?), and guessed the new renderer would also allow for fixed map sizes’ caps to be upped greatly, such as being able to handle said fixed maps’ sizes made beyond the size ranges of 100,000s to perhaps, possibly(just as food for thought), over 1,000,000s, for example.

I mean, coincidentally, depending on how well they gradually adapt to them, it could be helpful for any game to go over the top on map sizes and challenge players of all kinds to complete them, especially speedrunners, be it platformer, for those still loving the “Point A to Point B” type of end-level goals, Metroidvania(well, their maps are already growing larger, anyway), or otherwise.

I could imagine players trying to conquer more larger levels, regardless of size or map amounts in a whole stage, with their own clever strategies…

Eh. I love seeing more bigger maps in 2D games, anyway. Been through them, conquered them, can’t wait for more.

5

No, actually there is effectively no limit on the size of fixed size maps and internally they are stored the same way as infinite maps (avoiding memory usage for empty chunks).

The only limit is on the size of the number you can type into the input field (currently 99999x99999), which is mainly to keep users from creating maps so large that they are not practical. If somebody really needs a bigger map, it can be easily created through the scripting API, but some tools (like Bucket Fill) are better avoided on such large maps and the file size will become an issue if tile layers are not stored in chunks (like is done for “infinite” maps).

Anyway, the problem with creating bigger maps in 2D games isn’t tooling limitation. The problem is filling such a map with interesting content.