Print

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

[uebersoldat]

Toady's FPS response really made me sad. It's the one thing preventing me from enjoying this game as I like to have an established fortress but I can't be bothered to play a game at 5-15fps. It's certainly a complaint from many of my friends so I'm not sure how there hasn't been many official complaints unless folks are just scared of complaining about it.

Seems about the only major problem with this amazing game. Of course, Toady owes us nothing, and sure has given us something so I'm thankful for what we have. I just wish I could enjoy a larger or longer-lived fortress.

What about it made you sad? What I got from it is that he hasn't noticed a significant increase in complaints (not that there haven't been many), besides a few issues introduced in 0.44 which he fixed, and he's aware of other improvements that could be made without a major rewrite (i.e. they're not too hard), and it's an ongoing process, but he doesn't have specific plans to make specific changes immediately.

Well my fellow Urist, just the overall ignoring of the FPS elephant in the room over the years while we blaze on with awesome new features. Logically to me, the more you add decreases the life of a fort little by little before the FPS monster shows up. I mean, it can't help right? So at best it's staying leveled out but for how long? I feel like there may be a point where our bubble pops and we can't progress unless there is a major re-write. It's scary!

I think that we just keep hoping for some massive leap in single-core performance that will negate all this one day...but who knows? In the meantime I still enjoy the peace and tranquility of a fresh embark for a few years then just retire the fort when my frames dip into the 20's and below.

[Putnam]

I doubt we can progress in any reasonable way without a major rewrite now.

[Urlance Woolsbane]

I doubt we can progress in any reasonable way without a major rewrite now.

This, but to prevent ancient code from gumming up the works. Obviously, Toady cannot be expected to remember every facet of his colossus, but I'd hate for the sort of problems he's had with raids to become the norm.

[Random_Dragon]

Yes. From what I've seen, code maintenance is something you really can't put off for prolonged periods of time. The longer you do so, the more annoying it will be to deal with later. Periodically patching up a few holes in there tends to mean less work in the long run.

[KittyTac]

I can play with 5 FPS.

[MrWiggles]

How do we know ToadyOne isnt doing code maintance? He hasnt really gone too deep in his methodology. No one but Bothers Tarn and maybe the person that did the SDL code have seen the source code.

[iceball3]

It's certainly not something that's bothered me, much. Though, i play with my framerate capped specifically at 40FPS, because it feels like it lets me appreciate just how much is going on at any given moment.
Amusingly, in a 7 dwarf fort on a 2x2 embark, I've managed to pull 161,000 FPS, which tanks to 600 when leaves are falling from trees. I wonder if there's a way to toggle that off to save some frames for those who are scraping for them?

[Inarius]

I always smash carefuly everything expandable, and i limit my fort to 40-50 dwarves, so I haven't any FPS issues, but it's a real problem in general, true.

[Robsoie]

Hopefully that elephant in the room will be adressed before it will reach a point in which it will become impossible to do so without rewriting the whole game all over again.

[PlumpHelmetMan]

Isn't mythgen supposed to majorly rewrite many aspects of the game? If we're lucky, FPS will be one of the things addressed.

[thvaz]

Isn't mythgen supposed to majorly rewrite many aspects of the game? If we're lucky, FPS will be one of the things addressed.

Unlikely, as the main culprits of low performance has nothing to do with world generation and world events and is mostly tied to pathfinding, temperature calculations and tracking of itens.

[MrWiggles]

Isn't mythgen supposed to majorly rewrite many aspects of the game? If we're lucky, FPS will be one of the things addressed.

And the thing that results in evetual fps cant be fix. You eventually reveal to many tiles, you eventually have to many items.
Unlikely, as the main culprits of low performance has nothing to do with world generation and world events and is mostly tied to pathfinding, temperature calculations and tracking of itens.

[Inarius]

I don't really know how all of this can be bettered, but I agree these 3 items are the most important as far as FPS is concerned.
However, I'm not sure FPS is the main issue for now. Quest system, in general, and everything around is in a much much worse state, in my opinion.

[PatrikLundell]

Revealed tiles should have very little to do with game FPS, as the game mechanics apply to those tiles regardless of whether they're revealed or not. Display performance is affected by the amount of stuff DF has to draw, so unrevealed tiles where the view is tends to speed things up. However, the surface is fully revealed (with a lot of greenery on it) so that should serve as a display performance baseline.
The number of items in view affects display performance, but there is nothing that says the implementation has to evaluate every item in the fortress periodically regardless of whether they're subject to anything or not: a lot of items should need to be evaluated only if something happens to them, such as flow or magma affecting the tiles they are on, for instance. However, that does not mean managing an administrative process to shift items between different lists/queues as things happen to them is better than just going through them all: it would have to be evaluated (or at least estimated). However, optimizations ought to be aimed at mature mid sized -> large fortresses rather than small early ones and that's a tricky target as it takes time to set up the test environment. I guess a lab fortress (or possibly multiple player supplied ones) with a set of tools to reshuffle internal structures to check the outcome of different strategies might be useful, but that's a fair bit of work.

[Putnam]

the game already checks items less and less often as they continue to sit idle (iirc it was 1 tick->2 ticks->4 ticks->8 ticks etc. up to 1024)