Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Danjen]

Hi guys,

I was trying to model simple water flow for a little game project of mine (more of a programming exercise than anything), but I wasn't having much luck with it. So, I figured I'd try my luck here, since DF has a pretty solid fluid system in place, and you guys probably know the quirks of it better than I do.

Now, I have done some research on DF fluids, so don't worry. The wiki wasn't incredibly useful for specifics, but a few posts by Kanddak were very helpful. In order to emulate it though, I need to understand it...

Spoiler: 2x2 grid (click to show/hide)

I went into the object arena, paused the game, and put down a 2x2 grid of water. I stepped frame-by-frame until one of the water tiles updated, noting how many frames it took it to update. I stopped once all the initial water had moved. Now as you can see, 1 frame does not necessarily indicate 1 update. It takes an "unpredictable" number of turns before a fluid will move. Does each tile have an "update timer" similar to a creature's speed? Is this related to fluid density (or whatever the property is called) in the raws?

It's the only explanation I can think of for the "staggered" updates, both in regards to tile position and time, especially since I placed all the water on the same frame on a dry area. Take a look here for a more exaggerated example:

Spoiler: 4x4 grid (click to show/hide)

However it seems that whenever a tile DOES update, it will go down to 1/7, dispersing everything else to its neighbours, or at least in the case of a flat area with only minimal water present.

[hermes]

In a YouTube interview Toady gave recently, IIRC, he talked about spreading processes out over a 100 tick cycle, or something like that, to avoid bottlenecks that cause stuttering.  This sounds connected, possibly.  Can't find the link right now.  I guess that would explain the different decay times on the piles of water?

[Danjen]

Yes, that could be it. I suppose there might be a simple chunking going on under the hood. That gives me some insight into it.

[Boltgun]

Fluidity has an effect, magma take slower to flow despite using the same mechanics. Perhaps there is a cap of the number of tiles that are checked too, to reduce the FPS loss.

[Sizik]

Also, if you have a bunch of 7/7 water with a few 6/7 tiles in it, you can see a definite pattern in the way they move around.

[hermes]

Here is the interview I was thinking of, from about 21:00 onwards he talks about optimisation.  22:00 has the part about "staggered timers" and fluids.