Print

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

[Borge]

I'll ask the specific question. It seems one of the biggest causes of FPS death becoming such a considerable problem in DF2014 is the simulation of the game world during fortress play. In 34.11 after 20 years of play in a labyrinth of a 5x5 fort i am still getting 56 FPS, while after 4 years with 190 dwarves and 1 cavern open on a 3x3 embark in 40.11 I get 14 FPS. Is the simulation of the world worth the price we currently have to pay? And can you ever seeing this issue being resolved while maintaining world simulation? At the moment i just don't see it as being worth it as it doesn't seem to have a major impact on fortress play beyond hearing news.

[StagnantSoul]

The news is normally about how the goblins have slaughtered everyone, and there's no way for you to win- until you kill them off with three tough dwarves. If it weren't so broken, and you got world simulation from say... Twenty map spaces away+ Mountainhome? That may be more reasonable. I checked legends, I was getting constant reports of a goblin army that literally could never get to me.

[bluephoenix]

Well, it doesn't seem like it now but it will be worth it further down the road.
Once we can gather an army and attack other sites, have our own caravans trade with other civs and all sorts of other good stuff, we need to have an activated world.

DF 34.11 is still available to play if you want to play without the new features of 40.xx.

[Borge]

Well, it doesn't seem like it now but it will be worth it further down the road.
Once we can gather an army and attack other sites, have our own caravans trade with other civs and all sorts of other good stuff, we need to have an activated world.

DF 34.11 is still available to play if you want to play without the new features of 40.xx.

With all the bugs of the old version also including invulnerable sponges. I dare say those features of attacking other civilizations are a long way off, maybe in DF2016. Instead there should be a way to limit or halt world simulation through options.

[StagnantSoul]

Yeah, maybe something in the init files. It is rather annoying, and I've stopped my modding and my long term goal of making a fortress full of demon slayers, because the FPS drop is just killing my moods all the time.

[Meph]

It part of the reason I havent updated my mod yet. Still waiting for more updates, which hopefully fix more issues like that.

[StagnantSoul]

Player cancels mood: could not find good FPS.

[BesorgterZwerg]

I simulated a small world and embarked in a desert. Hopefully i get higher fps.
Anyway, i think that DF just needs multicore support to run fine.

[mifki]

Yep, I wonder what Toady is going to do with this. Parallelization of a complex algorithm can be hard or even impossible, but it has to be done at some point, otherwise this problem won't resolve by itself.

[Goatmaan]

I'm no expert just a pc gamer since 1985.

With all the unimplimented/unfinished features. ALREADY
In the game bringing fps death, this game will have serious problems
Adding more and more. Optimization can only go so far.
The bigger it gets, the harder it'll be to take multicore.
Right now WE are doing optimizations by small worlds short histories, and so
On.
Pretty sure FPS DEATH is here to stay, until DF goes multicore.

Are single core processors top sellers still?
Will they EVER BE again?

  Goatmaan

[smjjames]

I don't think he has gotten to the optimization phase yet.

[mifki]

I don't think he has gotten to the optimization phase yet.

Sure but 1st, making something to use multiple cores may be almost the same as rewriting it in some cases (especially taking into account the number of global structures in DF), so the sooner the better; and 2nd, maybe he won't get to optimisation phase in the next 15 years, in this case there's a danger of it becoming a beautiful but unusable world sim instead of a playable game.

[smjjames]

The whole 'when will DF use multithreading' (and variants of) topic has come up NUMEROUS times, so, it's not the first time the topic has been raised.

And I meant the optimization phase of this version, not in the sense of it being a development phase after alpha.

Anybody know what the FPS difference between the initial 0.31/0.34 release and after optimization was?

[GenericOverusedName]

Multithreading is a huge pain in the ass to do even when you set out explicitly planning to do it. Trying to break up a single-threaded thing into multiple threads is even worse...

I've found that having smaller worlds helps, as does setting the CPU priority higher if you can. If you can possibly just dedicate a core to the program exclusively, do that. Make sure you're not running out of RAM or anything like that as well.

The stuff that worked to help FPS in prior versions for the most part are still relevant now. Don't have a billion baby animals pathing everywhere, avoid having too much junk laying around, avoid having huge liquid flows and stuff. Stockpile streamlining and setting up hauling routes can help a lot by reducing the pathing calculations. Instead of having some massive stone stockpile next to your masons that require dwarves to haul stuff across the entire fort one rock at a time, set up a stockpile near your source of stone and set up a  hauling route from that source to a smaller stockpile near your mason workshops. Set the mason pile to take from the source pile exclusively. A bucket-brigade sort of hauling system works well (shorter distances = shorter pathing calculations), but setting up a track and minecart is great particularly for when you are moving stuff very long ways. Intermediate stockpiles and hand-hauling is just fine for things like wood hauling from the surface, usually.

Setting up a minecart and track system means you can have just one dwarf guiding a cart across your fort and back for every load of N stones, instead of having N dwarves hauling N stones by walking back and forth and back and forth... you don't NEED to have the tracks powered with rollers, a dwarf can guide a cart uphill just fine. If you want to automate it by all means feel free to do so, but it isn't explicitly necessary.

Turning off temperature and weather has a HUGE effect on FPS for me, but that's at the expense of not having temperature and weather. I'll drop a good 40 FPS when it starts raining in my fort. It's up to you to decide if you want those things or not. Personally, I really like having the weather and temperature so I just grin and bear the framerate hit. To each their own. You can change the settings without needing to regenerate the world (you will need to save, exit, and reload, though) so see if it helps for you.

[Footkerchief]

As far as I know, there's zero evidence that the new world simulation has a major impact on FPS.  However, there are many new and old bugs (a few of which have already been fixed) that do have a demonstrable major impact on FPS -- the usual stuff like pathfinding spam.