Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 ... 200 201 [202] 203 204 ... 211

Author Topic: Future of the Fortress  (Read 767674 times)

Encrtia

  • Bay Watcher
    • View Profile
Re: Future of the Fortress
« Reply #3015 on: January 25, 2018, 04:24:57 am »

If someone offered to pay you enough money, would you drop everything and work on getting DF multithreading support?
ToadyOne has spoken about turning down such things in the past. He has also talked about how multithreading probably wouldnt improve DF very much, but hes not very well versed in it.

Gosh darn it. Could you provide citation for that? The multi-threading aspect that is.
I'd join your contribution Pootis :P was really wanting to ask the same question...
« Last Edit: January 25, 2018, 07:07:03 pm by Encrtia »
Logged
~When death smiles at you, just smile one back.

MrWiggles

  • Bay Watcher
  • Doubt Everything
    • View Profile
Re: Future of the Fortress
« Reply #3016 on: January 25, 2018, 04:52:50 am »

That has never been a strong suit of mine, There was a fourmite who was just fucking devlish with their ability to link and search through the thread. Its just a common question, with the same answer. But I'll try.
Logged
Doesn't like running from bears = clearly isn't an Eastern European
I'm Making a Mush! Navitas: City Limits ~ Inspired by Dresden Files and SCP.
http://www.bay12forums.com/smf/index.php?topic=113699.msg3470055#msg3470055
http://www.tf2items.com/id/MisterWigggles666#

iceball3

  • Bay Watcher
  • Miaou~
    • View Profile
    • My DA
Re: Future of the Fortress
« Reply #3017 on: January 25, 2018, 07:01:34 am »

If someone offered to pay you enough money, would you drop everything and work on getting DF multithreading support?
ToadyOne has spoken about turning down such things in the past. He has also talked about how multithreading probably wouldnt improve DF very much, but hes not very well versed in it.
Gosh darn it. Could you provide citation for that? The multi-threading aspect that is.
I'd join your contribution MrWiggles :P was really wanting to ask the same question...
Multi threading is a very design-specific decision. Unless you know how it actually works, how DF actually works, and the shortcomings of the choice, then suggesting the change is vain.
https://en.wikipedia.org/wiki/Thread_(computing)#Multithreading - A short primer.

Currently, one instances where I think multithreading -could- potentially help is in worldgen. I can't say anything for actually making worldgen faster, but having a thread that accepts user input would be great for canceling worldgen where frame ticks take more than three minutes to advance. Without such an accomodation, the current behavior is that  windows thinks that DF is not responding anymore between ticks, and the trying squeeze in they keypress input needed to end worldgen becomes increasingly difficult.

KittyTac

  • Bay Watcher
  • Impending Catsplosion. [PREFSTRING:aloofness]
    • View Profile
Re: Future of the Fortress
« Reply #3018 on: January 25, 2018, 07:09:29 am »

If someone offered to pay you enough money, would you drop everything and work on getting DF multithreading support?
ToadyOne has spoken about turning down such things in the past. He has also talked about how multithreading probably wouldnt improve DF very much, but hes not very well versed in it.
Gosh darn it. Could you provide citation for that? The multi-threading aspect that is.
I'd join your contribution MrWiggles :P was really wanting to ask the same question...
Multi threading is a very design-specific decision. Unless you know how it actually works, how DF actually works, and the shortcomings of the choice, then suggesting the change is vain.
https://en.wikipedia.org/wiki/Thread_(computing)#Multithreading - A short primer.

Currently, one instances where I think multithreading -could- potentially help is in worldgen. I can't say anything for actually making worldgen faster, but having a thread that accepts user input would be great for canceling worldgen where frame ticks take more than three minutes to advance. Without such an accomodation, the current behavior is that  windows thinks that DF is not responding anymore between ticks, and the trying squeeze in they keypress input needed to end worldgen becomes increasingly difficult.

That would be useful.
Logged
Don't trust this toaster that much, it could be a villain in disguise.
Did you know that cats that have mysteriously disappeared from home can come back after a few months without warning?

Inarius

  • Bay Watcher
    • View Profile
Re: Future of the Fortress
« Reply #3019 on: January 25, 2018, 07:48:44 am »

Quote
Currently, one instances where I think multithreading -could- potentially help is in worldgen.
I Agree.
Logged

Encrtia

  • Bay Watcher
    • View Profile
Re: Future of the Fortress
« Reply #3020 on: January 25, 2018, 12:13:25 pm »

Multi threading is a very design-specific decision. Unless you know how it actually works, how DF actually works, and the shortcomings of the choice, then suggesting the change is vain.
https://en.wikipedia.org/wiki/Thread_(computing)#Multithreading - A short primer.

By your misreading of what I typed as an assumption of me suggesting a change, I'm content with moving on.

Instead, let me ask, how do you reckon one might circumvent the FPS drain beyond limiting the game's potential as a player?
Logged
~When death smiles at you, just smile one back.

Knight Otu

  • Bay Watcher
  • ☺4[
    • View Profile
Re: Future of the Fortress
« Reply #3021 on: January 25, 2018, 01:35:46 pm »

Given the current state of how dwarves can currently cope with stress extremely well in the long term, how do you see tweaking the stress system going forward?
Is it in it's current state:
-Intentionally, the current dwarven stress resistance (and getting to -999999 stress) with little attention is serving as a baseline.
-On accident, was tuned to an unintentional degree.
-Or in preparation for more intense and common stressors being available in future updates, or a different scaling system upcoming, or similar overhauls?
I am just wondering, because I've been pondering over a mod that makes dwarves more difficult to keep unstressed, and it's really hard to balance with how stress currently works, so I'm just wondering what paradigm you're considering for how dwarves can and should cope with stress.


Stress is likely to get some attention during this bugfix cycle.
« Last Edit: January 25, 2018, 01:41:18 pm by Knight Otu »
Logged
Direforged Mod v1.3b for 0.43.05 (Updated Dec 28th 2016)
Random Raw Scripts - Randomly generated Beasts , Vermin, Hags, Vampires, and Civilizations for 0.42.06 (Updated Mar 27th 2016)
Castle Otu

iceball3

  • Bay Watcher
  • Miaou~
    • View Profile
    • My DA
Re: Future of the Fortress
« Reply #3022 on: January 25, 2018, 03:25:26 pm »

By your misreading of what I typed as an assumption of me suggesting a change, I'm content with moving on.

Instead, let me ask, how do you reckon one might circumvent the FPS drain beyond limiting the game's potential as a player?
Current consensus as of recent is that, while processor power helps, another big help is upgrading your RAM, toward improving the frequencies and latency of the memory you'll use, largely due to the fact that DF is constantly in a state of reading and accessing ram addresses, which can bottleneck if your CPU is up to the task. Having enough RAM to begin with isn't difficult, but if you don't have enough ram, that can severely hamper performance too.
Ingame, a lot of not-too limiting things can involve cramming as many of your non-grazing non important and baby animals into cages as possible. Using traffic designations can also optimize pathfinding as well for instance, if you designate the doorways of multi-doorway rooms as low traffic, dwarves will be less likely to try to consider a "shortcut" through it, saving on a few processing ticks here and there. Blocking off large unused subterranean areas with forbidden doors will also help reduce the amount of pathfinding that stretches to unneeded areas. That's just a few pieces of advice, anyway. Also apologies if my initial comment were a bit scathing, it honestly largely concerned the original post concerning "How much money would toady accept before he focuses only on x feature" question.
Stress is likely to get some attention during this bugfix cycle.
Ahh, that's very good, and definitely answers my question. Thank you!

Putnam

  • Bay Watcher
  • DAT WIZARD
    • View Profile
Re: Future of the Fortress
« Reply #3023 on: January 25, 2018, 03:56:32 pm »

Off the top of my head, fluid calculations could absolutely get help from multithreading. Keep track of separate rivers/seas/ponds etc., calculate fluid flows in them concurrently. There would be some overhead from combining multiple water areas, but that doesn't happen nearly as often as within-area flows. There might also be overhead from expanding the area a particular flow could get into, which is a bit ugly, and determining exactly how far-reaching the flow ought to be might also cause problems.

Yeah, even the simplest optimizations end up being tough.

Encrtia

  • Bay Watcher
    • View Profile
Re: Future of the Fortress
« Reply #3024 on: January 25, 2018, 07:00:20 pm »

Current consensus as of recent is that, while processor power helps, another big help is upgrading your RAM, toward improving the frequencies and latency of the memory you'll use, largely due to the fact that DF is constantly in a state of reading and accessing ram addresses, which can bottleneck if your CPU is up to the task. Having enough RAM to begin with isn't difficult, but if you don't have enough ram, that can severely hamper performance too.
Ingame, a lot of not-too limiting things can involve cramming as many of your non-grazing non important and baby animals into cages as possible. Using traffic designations can also optimize pathfinding as well for instance, if you designate the doorways of multi-doorway rooms as low traffic, dwarves will be less likely to try to consider a "shortcut" through it, saving on a few processing ticks here and there. Blocking off large unused subterranean areas with forbidden doors will also help reduce the amount of pathfinding that stretches to unneeded areas. That's just a few pieces of advice, anyway. Also apologies if my initial comment were a bit scathing, it honestly largely concerned the original post concerning "How much money would toady accept before he focuses only on x feature" question.

Oh, thank you :) Even if you didn't need to apologize, I really appreciated that!

I would say that everything you've typed is something I shall pass on due to its usefulness, though I already utilize such techniques to the best of my ability. However, I purposefully phrased the question with "beyond limiting the game's potential as a player" to highlight that I, as the player, am forced to apply inhibiting techniques (as generously provided), compromising my freedom (not talking mega-projects) when playing the game. Without going into extraneous detail, it's simply go-to advice for players (on any PC, as we're aware) to avoid whole sections of the game (liquids, 3x3 or larger embarks, HFS, caverns even) just to maintain acceptable frames per second (being ideally the starting 100fps speed imposed by the game, but really staying above 35fps at the moment). Consequently, it was more of a "what can be done from the Developer's side to alleviate such issues if Multi-Threading isn't the holy grail?"

My angle though, was wanting to see Toady's stance on the matter as explained by MrWiggle. There are so many opinions for & against multi-threading, that it's utterly pointless debating it with other users. It is, however, the main & most controversial optimization technique to cure what I, perhaps in the minority, consider the single biggest deterrent to this amazing game, being the FPS drag. (bugs don't count :P)

Again, thanks for the advice. In case you're wondering, below is how I've tackled it as the player. If I you bother to read it, would be grateful for any other techniques ???

Spoiler (click to show/hide)
« Last Edit: January 25, 2018, 07:47:33 pm by Encrtia »
Logged
~When death smiles at you, just smile one back.

KittyTac

  • Bay Watcher
  • Impending Catsplosion. [PREFSTRING:aloofness]
    • View Profile
Re: Future of the Fortress
« Reply #3025 on: January 25, 2018, 08:35:18 pm »

Playing with temperature off is not !!FUN!!. I always play with it on, I don't care if I get 10 FPS.
Logged
Don't trust this toaster that much, it could be a villain in disguise.
Did you know that cats that have mysteriously disappeared from home can come back after a few months without warning?

Encrtia

  • Bay Watcher
    • View Profile
Re: Future of the Fortress
« Reply #3026 on: January 25, 2018, 08:38:14 pm »

And now you see my predicament :P I'm having to loose out on !!FUN!! to attempt to gain more frames per second.
Logged
~When death smiles at you, just smile one back.

iceball3

  • Bay Watcher
  • Miaou~
    • View Profile
    • My DA
Re: Future of the Fortress
« Reply #3027 on: January 25, 2018, 09:18:49 pm »

Oh, thank you :) Even if you didn't need to apologize, I really appreciated that!

I would say that everything you've typed is something I shall pass on due to its usefulness, though I already utilize such techniques to the best of my ability. However, I purposefully phrased the question with "beyond limiting the game's potential as a player" to highlight that I, as the player, am forced to apply inhibiting techniques (as generously provided), compromising my freedom (not talking mega-projects) when playing the game. Without going into extraneous detail, it's simply go-to advice for players (on any PC, as we're aware) to avoid whole sections of the game (liquids, 3x3 or larger embarks, HFS, caverns even) just to maintain acceptable frames per second (being ideally the starting 100fps speed imposed by the game, but really staying above 35fps at the moment). Consequently, it was more of a "what can be done from the Developer's side to alleviate such issues if Multi-Threading isn't the holy grail?"

My angle though, was wanting to see Toady's stance on the matter as explained by MrWiggle. There are so many opinions for & against multi-threading, that it's utterly pointless debating it with other users. It is, however, the main & most controversial optimization technique to cure what I, perhaps in the minority, consider the single biggest deterrent to this amazing game, being the FPS drag. (bugs don't count :P)

Again, thanks for the advice. In case you're wondering, below is how I've tackled it as the player. If I you bother to read it, would be grateful for any other techniques ???

Spoiler (click to show/hide)
32 GB, that's quite a lot. A lot more than important (once you have "enough", more than enough isn't a big deal). What's the Hz frequency and the CAS latency?
Also out of curiosity, what's the population?
« Last Edit: January 25, 2018, 09:24:04 pm by iceball3 »
Logged

KittyTac

  • Bay Watcher
  • Impending Catsplosion. [PREFSTRING:aloofness]
    • View Profile
Re: Future of the Fortress
« Reply #3028 on: January 25, 2018, 09:23:15 pm »

And now you see my predicament :P I'm having to loose out on !!FUN!! to attempt to gain more frames per second.

I do not care about FPS.
Logged
Don't trust this toaster that much, it could be a villain in disguise.
Did you know that cats that have mysteriously disappeared from home can come back after a few months without warning?

iceball3

  • Bay Watcher
  • Miaou~
    • View Profile
    • My DA
Re: Future of the Fortress
« Reply #3029 on: January 25, 2018, 09:25:02 pm »

And now you see my predicament :P I'm having to loose out on !!FUN!! to attempt to gain more frames per second.

I do not care about FPS.
It can be important, in a way, but it largely depends on one's acclimation. I've limited my max fps to 40 from the very start, and I'm used to it to the point where I can pull rather  decently populated fortresses with little slowdown from my comfort zone.
Pages: 1 ... 200 201 [202] 203 204 ... 211