Print

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

[Urist McDepravity]

Something that's worth bearing in mind here is that the game is currently less than one third feature-complete. At this point Toady One's approach seems to be to put most of his effort into getting the content to fit his vision of the game DF is intended to become, only changing code he's already written when absolutely necessary. Systematic streamlining and optimisation will probably take place after we hit v1.0.
This may be frustrating for us in the short term, but it will ultimately bring us the finished product a lot faster.

Yeah, as software engineer myself I know his approach is mostly correct, 'premature optimization is the root of all evil' (c).
Besides, optimizing code which is going to be rewritten later is a waste of time anyway.
But it sucks being in player's shoes nevertheless.

[Zaranthan]

It's been asked several times, but not answered: why are ramps faster than stairs?

A staircase has ten tiles connected to it: the eight surrounding it horizontally, plus the tiles directly above and below.

A ramp has eight tiles connected to it: up to seven horizontally, plus the tile above the adjacent wall.

Combine this with the fact that most people who use stairs have massive shafts, whereas ramps actually force you to restrict where pathfinding can go, and you get much better performance by designing your fort with ramps.

[scira]

On slightly less than normal embark sizes i can get  about 100 fps to start which goes to about 20 by about 100 dorfs, trying to play on a 16x16 gives about 25 fps at start, a little more if i kill 6 of the dwarfs.
Now I play on 2x2's and its almost a perfect setup, my cap of 500 fps at start goes down only to about 90 by the 100 dorf mark. Plus it kind of forces me to try to optimize my fort for the long term, something i like.
This is 40d of course, I like hunting, fishing, killing things and not crashing.

[Hyndis]

It's been asked several times, but not answered: why are ramps faster than stairs?

A staircase has ten tiles connected to it: the eight surrounding it horizontally, plus the tiles directly above and below.

A ramp has eight tiles connected to it: up to seven horizontally, plus the tile above the adjacent wall.

Combine this with the fact that most people who use stairs have massive shafts, whereas ramps actually force you to restrict where pathfinding can go, and you get much better performance by designing your fort with ramps.

What about using stairs in the manner of using ramps? For example, my stairways are grand, roomy, and move diagonally through the rock.

Code: [Select]

XXXX..XXXX
X........X
X.<....<.X
X.<....<.X
X........X
XXXXXXXXXX

Up stairs are on the right, down stairs are on the left. I do not use up/down stairs except for very rarely used maintenance for some of my complex machinery. Each level of the stairs has a landing. Preferably with masterwork statues installed in the room, and double doors leading out of the landing to the level the stairs go to for other rooms and corridors.

Using ramps would likely have about the same effect.

[Richards]

The temp calcs are pretty big.  I tested it on a 120 dwarf fort that I was getting 30 FPS, turned it off, jumped up to 50.  Theres something wrong with that.

Ditto. There's definite code optimization missing. Toady even remarked something along the lines that he's not satisfied with his code as it is. I think Baughn's work on the display code is a good foundation on the fact that Toady could pass out some work on Dwarf fortress. Namely pathfinding. I think if he could find some trustworthy coders it'd be beneficial for it to be worked on by others so Toady can focus elsewhere. As it is, DF needs it.

Also from what I know about coding, though DF is a complex game most of the calculations done for things like fighting, item making, etc, are simple, they don't tax the CPU. Even though a game may be complex for you that doesn't mean it's complex for the computer too.

[Jake]

%   Release Date
21   8/9/2006
22   10/2/2006
23   1/18/2007
27   10/29/2007
28   7/16/2008
31   4/1/2010

Based on a linear trendline, the v1 is projected to be here sometime around 2034.  I think it is amazing what Toady has done.  Lets just be realistic on how long this ambitious design will take to complete.  Putting off optimizations for another 24 years just isn't going to work.
[/quote]
I didn't say I thought he's putting it off completely, let alone that he should be, just that it's being done ad-hoc whilst he either codes in new features for the next release or fixes bugs in the most recent. I daresay the code is currently an ugly, hacky mess that's using up a lot more CPU cycles than it ought, but there's no point going in and honing it to perfection when there are immediate plans to add a bunch of new functions that are inevitably going to break something.

This is a problem we wouldn't have if Toady would let someone else give him a hand with this stuff, though I don't suppose there's a cat in Boatmurdered's chance of that after the whole Khazad fiasco. Unless he wants a Vice-President (IP Protection) who owns his own axe handle and is willing to travel?

[goffrie]

Ahem. So apparently I'm dumb. Yeah, DF doesn't use manhattan distance. However, it does not use a euclidean distance metric either.

[Urist McDepravity]

though I don't suppose there's a cat in Boatmurdered's chance of that after the whole Khazad fiasco.

On the contrary, with that binary slipped, its dead-easy to reverse-engineer it back into sources and and fork it, so there's not that much logic in being protective over sources anymore.
But thats why IP laws exist. And thats why non-free open-source is possible as well. Toady could always release bits he needs help with under over-restrictive license, explicitly prohibiting derivative works, re-distribution and inclusion in other works.
Sure that could not stop some individuals from making rogue forks and even trying redistribute them, but that would prohibit them getting popular and/or commercial.

[Richards]

though I don't suppose there's a cat in Boatmurdered's chance of that after the whole Khazad fiasco.

What exactly happened with Khazad?

[Urist McDepravity]

What exactly happened with Khazad?

http://www.bay12forums.com/smf/index.php?topic=58796.msg1310378#msg1310378

[zwei]

though I don't suppose there's a cat in Boatmurdered's chance of that after the whole Khazad fiasco.

What exactly happened with Khazad?

Someone decided to make DF clone from scratch by expanding one visualizer. Insane goal.

As such, it is of no consequence to current discussion as he had no aid or souce from Creator. It pretty much happend without any Creators input.

In fact, Baughns help with graphics engine was only (was it?) such cooperation and it was actually succesfull.

[Urist McDepravity]

As such, it is of no consequence to current discussion as he had no aid or souce from Creator. It pretty much happend without any Creators input.

Not exactly, you missed 'binary with symbols' part. Its not sources, but still quite easy to reverse-engineer.

[Lemunde]

Heh, for the longest I was actually running with 30 fps as the max speed.  I preferred being able to see exactly what my dwarves were doing without them zipping across the screen.  I'm not quite as patient now but I might go ahead and lower it back down to 50 if only to fool myself into thinking the game isn't running slow.

[Gnoll Fortress]

Am I the only one who doesn't have massive fps issues when playing my processor is 3.2 ghz like the op, but I never reached 20 fps even with 200 dwarfs on a 5x5 site(at 200 dwarfs its more like 40 fps). I still would like to see the code optimized so I can raise the population cap to 500.

I don't understand some of the drains on cpu people are talking about how can item count cause strain on the cpu aren't pcs basically designed to count things?

[expwnent]

Gnoll Fortress:

Interesting. There are a couple things that might possibly explain it. How much clutter do you have in your fort? What are your init options? How organized are you with the layout of your fort?

I agree that large item counts shouldn't have a big impact of framerate, as the program shouldn't have to do stuff with every single item on a frame-by-frame basis, but that doesn't mean that they won't have an impact. It could even be more subtle. If your program has more memory usage, it can lead to bad cache performance, which makes stuff slow. Speaking of which, how much RAM do you have?