Print

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

[Schmaven]

I'm not sure about this or not, but I wonder if less activity happens in the caverns before they are discovered compared to after they have been breached?  I've seen a problem with a reanimating biome where some cavern creatures get into a fight, some die, then it catsplosions its way into hundreds of severed arms crawling around and masses of undead crundles stuck atop and between mushrooms growing at the map edges.  Easily a 10 FPS hit if 25+ years of that is left on its own.  So I have a theory that perhaps just world-gen style activity would occur in unrevealed caverns so long as they remain undiscovered.  Or maybe just the pathfinding calculations would not occur in that unrevealed glittering blackness?  I haven't been able to successfully avoid them for any significant length of time to test this.  If I were to test it, I think I'd have to generate a new world with lots of z-levels above the 1st cavern layer and just not go deeper than that for 50 or so years, then breach them and see what's going on and compare FPS averages.

[delphonso]

I think through stories like the sparrow person naming their spear a hundred times, there is certainly some activity in the caverns even if you don't see it. I believe forgotten beasts don't arrive until you get there, though.

Generally, forgotten beasts are probably helpful to FPS, if your caverns are small enough - because they'll just kill everything on the map for you. And as long as there is no possible path for your dwarves, it shouldn't cause pathing issues to recover the dead bodies or bits of silk or whatever.

[gchristopher]

4. USE A SMALL WORLD. 65x65 is the most you should be using. Use advanced params if you want more stuff, but keep it limited; my rule of thumb is that if you have to wait any more than 1 second per year generated, you'll notice some serious hit to in-game FPS. Part of why Archcrystal is so slow (and was even during its formative years) is because Sethanos used a massive world with a ton of population for no reason, since even now he's not acting on a global scale, merely a regional one. Seth also has a more powerful computer than average even with his bad FPS.

+1. World activation seems to create an ever-increasing drain on FPS, which is the main killer of long-term forts. The smaller the world, the better.

6. For defense, atom smash or magmatize goblins directly. Burning goblins generate a lot of lag, so killing them before dropping them into magma is helpful. It's actually pretty hard to make fully automated defenses, since even if you can kill them you still have to clean it up without input or civilians, and you need to account for things like creatures too big to be smushed and trapavoid creatures. Or just disable sieges.

I find that running out of interesting invaders is more of a problem than killing them, even after lowering the siege pop triggers, setting titan attack to 1 in world gen, and setting mega- and semi-mega beast attack triggers to 1. I recommend impulse ramp loops with gold or rose gold minecarts to grind up invasions. Use a bridge as an off-ramp to open it up for cleaning.

[NordicNooob]

I find that running out of interesting invaders is more of a problem than killing them, even after lowering the siege pop triggers, setting titan attack to 1 in world gen, and setting mega- and semi-mega beast attack triggers to 1. I recommend impulse ramp loops with gold or rose gold minecarts to grind up invasions. Use a bridge as an off-ramp to open it up for cleaning.

I've had a properly generated world get me pretty much infinite goblin invaders to the point I regularly kill sieges of several hundred and they don't ever dry up. Though yes, large beasts and necromancers do run dry eventually, and there's not much you can do about that.

I'd personally advise against a minecart crusher like that. It's incredibly effective, but requires manual cleaning, which is extremely problematic not just in long-term forts, but just small forts in general. A flamethrower is probably the best automatic defense I can think of (as it doesn't care about size, unlike a much simpler trap that just yeets stuff into magma, only fire immunity, and once large beasts run out you won't see anything fire immune on the surface), you should be able to have it run for a reasonable timeframe after a pressure plate press given some basic dwarfputing. If you just want to delete the remaining drops, you could use a falling edge signal when the flamethrowers turn off to flick a bridge open and shut to smush all the remaining junk.

[bool1989]

I gotta say, I'm leary of atom smashing stuff, cause there's no way that stuff isn't bug and/or chrash prone.

When I play DF it tends not to be very stable, and I'm already quantum stockpiling stone. I don't want to make my save super buggy and crash prone just cause I want to get rid of some old clothing.

[delphonso]

Atom smashing is the only true path.

(I don't think there are any bugs relating to atom smashing. There's even mechanisms in the code for dealing with atom-smashed artifacts.)

If you're looking for a long-term fort, speed of simulation is essential. Just make all your dwarves always nude, honestly I don't know if the mood debuff will be that bad.

[PatrikLundell]

Caravan garbage disposal works fairly well, and has the advantage of not hitting clothiers with masterwork destruction shock.
Magma disposal works for everything that isn't magma safe (but results in masterwork destruction shock if such items are destroyed).

I tend to export worn clothing and atom smash enemy corpses (this requires some micro management to disable the corpse atom smashing when dorfs die, as they may be hauled to the corpse stockpile before being buried, and corpse stockpiles have no detailed parameters to allow you to make one for fortress members and one for enemies).

[Quarque]

Just make all your dwarves always nude, honestly I don't know if the mood debuff will be that bad.

uh.. this bit is not a good idea, sorry to say it. Your dwarves will go insane. If you'd like to avoid cloth clutter, I think NordicNoob had the best advice. Go with armor when possible (so with the exception of woodcutters and kids), and use a refuse pile  to make them rot away quickly, or trade them away.

[Urist9876]

Another way to handle clothing is having a goods stockpile linked to your clothing and leather working shops and have it set to accept from links only and another stockpile for clothing that is combined with a refuse stockpile.

This way only new clothing will not rot away and used clothing will be discarded and rots away fast because of the refuse stockpile trait. Masterwork clothing that rots away does not seem to give bad thought.

Downside is you cannot buy clothes from caravans, as anything you do not make yourself will end up on the refuse stockpile. After a little dying and tailoring practice, non local made clothes are inferior anyway.

This might be not the best way for fps, but it can be set up in a way that requires very little attention.

By using an extra stockpile and overproduction of clothes, you can even set it up to only keep masterwork ones and have the others be transferred to the refuse stockpile (use one stockpile for the desired quality and another for low quality stuff that is linked to give goods to the refuse stockpile).

[pamelrabo]

There's this Longdeath Fort, a projected 1000 years fort, currently hundreds of years in, if memory serves.
Look for BlindiRL on twitch or Youtube, he streams regularly and maybe he can give you some good advice on stream or in his vod youtube.

https://www.twitch.tv/blindirl

[Putnam]

(this post was linked to on reddit, sorry)

4. USE A SMALL WORLD. 65x65 is the most you should be using. Use advanced params if you want more stuff, but keep it limited; my rule of thumb is that if you have to wait any more than 1 second per year generated, you'll notice some serious hit to in-game FPS. Part of why Archcrystal is so slow (and was even during its formative years) is because Sethanos used a massive world with a ton of population for no reason, since even now he's not acting on a global scale, merely a regional one. Seth also has a more powerful computer than average even with his bad FPS.

+1. World activation seems to create an ever-increasing drain on FPS, which is the main killer of long-term forts. The smaller the world, the better.

Sorry to say this is just unequivocally false; it does cause a bit of a drain, but it's a flat drain, the same no matter how long your fort goes. This is obviously true even when thought about logically, but world activities show no evidence in my profiling even on medium or large long-lasting worlds. It's negligible.

I gotta say, I'm leary of atom smashing stuff, cause there's no way that stuff isn't bug and/or chrash prone.

When I play DF it tends not to be very stable, and I'm already quantum stockpiling stone. I don't want to make my save super buggy and crash prone just cause I want to get rid of some old clothing.

DF is pretty stable, crashes are probably TWBT

Make sure you have a reliable way to replenish your population, whether it be via natural births, or reliable migrant waves.  For migrant waves, being able to successfully trade is pretty important.  So have a protected entry from whatever sieges / giant undead birds / thralling clouds / toxic rain that your embark might visit upon you.

I wouldn't recommend a balanced food / booze production though.  Stockpile it to the limit, it doesn't go bad.  To save on space, I usually have food and drink quantum stockpiles directly under the main food and drink stockpiles.  But locked away in their 1 tile room.  That way when calamity inevitably strikes, and you have to shift everyone into slab carving / corpse cleanup duty, you simply unlock the doors and don't have to worry about food for a couple years while you keep everyone sane and regain your footing.  Just disable the QSP loading mechanism if it's sucking out all your food and drinks too quickly, and toggle it back on briefly to deal with the bumper crops.  Yes, yes, FPS and whatnot related to large stockpiles...  Just cap it at some point.  Maybe run the fort for 50 years without it, then stock it up to 1,000 each and see if you even notice a difference.

A panic room, fully stocked with food, drink, soap, a well, medical supplies, a couple picks, good armor and a variety of weapons is also helpful if you intend on avoiding reclaim missions.  I suppose the alternative is to not ever let things get so bad where you need a panic room, but where's the fun in that?

I only do quantum stockpiles for stone, since stone is such a pain to stockpile normally.

For food and drink I set up repeating 'prepare easy meal' and 'brew drink from plant' orders with the appropreate conditions. With the right conditions everything is stored up in barrels or stone pots, sometimes clay or glass pots when I can get them.

With my special world gen parameters I have 5000 emark points to work with, so I bring a lot of drink, food, plants, and seeds with me, and I usually embark in places with soil, so a solid food industry usually isn't a problem.

I'll probably do the panic room thing, though.

lavish meals are simply superior in every way, including for FPS (it combines 4 stacks of items into one, and stacks are the things calculated on, not individual items)

I wouldn't recommend a balanced food / booze production though.  Stockpile it to the limit, it doesn't go bad.

this, however, is definitely no good (I was recently called out for screenshotting my having 1000 plants as having too many in a 30-dwarf fortress), FPS loss is caused by units first then items

[Salmeuk]

Sorry to say this is just unequivocally false; it does cause a bit of a drain, but it's a flat drain, the same no matter how long your fort goes. This is obviously true even when thought about logically, but world activities show no evidence in my profiling even on medium or large long-lasting worlds. It's negligible.

I believe the original quote was referring to the notion that longer, larger worlds produce less FPS in fortresses played on those worlds. The longer a world exists, the more population, at least up to the cap placed by init settings. that population causes lag. I would agree though that during play you are unlikely to experience a dramatic decrease in FPS over time due to this activated state. Population does not change enough in the ~30 or so years to affect the FPS.

If you still disagree, I want explanation!!1! because in all my experience with large, long-lived worlds, the most consistent experience was FPS drain occurring earlier in the fortresses life cycle

[Putnam]

Sorry to say this is just unequivocally false; it does cause a bit of a drain, but it's a flat drain, the same no matter how long your fort goes. This is obviously true even when thought about logically, but world activities show no evidence in my profiling even on medium or large long-lasting worlds. It's negligible.

I believe the original quote was referring to the notion that longer, larger worlds produce less FPS in fortresses played on those worlds. The longer a world exists, the more population, at least up to the cap placed by init settings. that population causes lag. I would agree though that during play you are unlikely to experience a dramatic decrease in FPS over time due to this activated state. Population does not change enough in the ~30 or so years to affect the FPS.

The wording of "ever-increasing drain on FPS, which is the main killer of long-term forts" definitely suggests "long-term forts are hit harder by world activation than short-term forts", which is the part I'm objecting to.

[☼Obsidian Short Sword☼]

Don't use animals. And if You really need eggs, then lock them up in a 1x1 room(Edit In reply to Putnam: But Don't use tightly closed doors.).

[Schmaven]

I wouldn't recommend a balanced food / booze production though.  Stockpile it to the limit, it doesn't go bad.

this, however, is definitely no good (I was recently called out for screenshotting my having 1000 plants as having too many in a 30-dwarf fortress), FPS loss is caused by units first then items

To quantify the effect of item count on FPS, I loaded up a recent fort with over 50,000 prepared meals, about 524 barrels of which are in a QSP, and let it run for a week or so for FPS to stabilize.  It bounces up and down a bit, but seems to mostly be around 32

I then spawned some magma on that tile using DFHack to destroy the QSP load.  Unfortunately, most of the meals were in green glass pots, so only about 20,000 were able to be destroyed in this simple way.  The smoke cleared, and I let it run for another few days for FPS to find a new equilibrium while viewing the same part of the fort, was interrupted by the game pausing as a dwarf went insane, and again as he claimed a workshop.  But the FPS readout at the top did not increase, rather it stayed fairly consistent at 32.

Assuming that the number of items does indeed negatively impact FPS in a substantial way, it appears my particular fort has long ago crossed some threshold where that negative impact has reached a limit whereby further increasing the item count no longer matters to FPS.  There are excesses in every stock category, so destroying 20,000 may not have been a relatively large enough reduction to see any effect on the overall item calculation load.  Given the current state of this particular fort, amassing prepared meals does not appear to effect FPS.  Reducing meals from 50,000 to 30,000 had no measurable effect anyway.

Ammo is by far the biggest line item in the stocks menu at 105,727 (only 5,328 stones due to a very lucrative mechanism trading scheme), so some more magma was spawned, and ammo was reduced to 22,217.  The fort was left to run for some more days, an artifact steel warhammer was created, a buzzard monster was riddled with 5 pages of marksdwarf arrows, and again, no effect on FPS was observed.  Based on this test, in my wildly unstable and inconsistent fort, the number of items appears to have little to no impact on FPS (100,000 total items melted, FPS constant at 32).  I suppose the next step would be to flood the whole fort with magma...  But well over 50% of the total item count in the stock menu has already been destroyed, so I wouldn't expect to see any significantly different results from that.

How much of an impact do you think item count has on FPS in general, and how did you come to that conclusion?

I suppose it could be an irreversible effect, and once items have been created, the FPS is thereafter penalized, no matter if that item is subsequently destroyed or allowed to remain.