Print

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

[Apani]

Simple idea: making "stockpile" buildings, that is buildings requiring lots of materials and made just to stuff them in one place and be deconstructed when needed.
Practic example:
Step 1: Have lotsa wood laying around
Step 2: build "Wood Pile"s from various logs near your workshops
Step 3: deconstruct one when you need wood, getting back logs it was made of.
Step 4: get killed (or worse, cluttered) by falling wood logs as you deconstruct the stockpile
Purposes: helping with FPS, helping with space management in aboveground settlements. I don't want to dig huge underground rooms like mices, cellars are fine but rarely sufficient.

[silentdeth]

Code: [Select]

###..
###S#
###..

# = Wood stockpile
S = trackstop set to dump to the east, with a minecart on top of it set to take from the 3x3 stockpile.
. = floor
Instant quantum stockpile.

[Apani]

I don't know if I can set that thing in my house...

[Meph]

Doesnt work, because even if I use 100 items for one workshop, the game still know that these 100 items are there. So no FPS improvement. Space, yes, but like mentioned this can be done with quantum stockpiling.

[silentdeth]

I don't know if I can set that thing in my house...

I find it amusing that you think a building Meph designed to do this would be smaller than that.

Quantum stockpiling does improve performance to a degree, not sure how much, but some from not having to draw all of the items.

[Meph]

but some from not having to draw all of the items.

I wouldnt be so sure about that, the total number of tiles the graphics have to paint is the same, it its a 16x16 boulder or a 16x16 floor tile.

[Apani]

I read on the wiki:
"Evidence has emerged that the quantity of items may not be the actual culprit, but rather the hauling, stockpiles, pathfinding and other CPU-intensive tasks they involve."

[Meph]

So the line of thought is: If the 100 wood logs are buildmats, then the pathfinding/AI ignores them, because they are no valid reagents. Thats reasonable in theory.

I still dont think that it has a big effect. You have read the Banners of Heart, so you know that getting rid of 6000 crafts only improved FPS by 1.

I mean, its not hard to write, I simple make a workshop that takes BUILD_ITEM:100:NONE:NONE:NONE and you can store any 100 items in it. No idea if that really helps, but someone (not me) has to to science on that. If people pack 1000 items into 10 storage workshops and it gives 1 FPS, then I wouldnt include it in the mod... not worth it. Mostly because of the micromanagent, if people need  a single of these items, they have to deconstruct the entire workshop, freeing all 100 items, scattering them around the construction site.

It also doesnt work on corpses, bodyparts, plants or food. Anything that rots will deconstruct the workshop when it rots away.

[Apani]

Talking about the Banners of Hearth, here are Jimbo's old observations:

I’m working on a project to optimize FPS. (This post is too long, apologies in advance.)

In vanilla, it’s reported that the top three reasons for abandoning a fortress are
1) boredom
2) FPS death and
3) unexpected  !FUN!

I’m thinking with Masterwork: boredom, not so much.  There’s lots of neat stuff to figure out and Meph keeps uploading new versions with new stuff …  but FPS is the trade-off and particularly so since many of us play on limited laptops.  So I’ve been studying the FPS thing and while the Wiki pages are good, they maybe could be updated for MW.  This isn’t far enough along to post as statistically valid but so far, Flow>Pathfinding>Stuff to count>Temp variations/Weather.  If charted, Flow and Pathfinding would make up by far the largest slices of the pie.  [I should probably stop here.  ]  What I’ve found so far:

Pathfinding: For this, the Wiki pages are pretty good.  A 2x2 embark is only 25% the area of a 4x4 and probably the single “best” plan (the trade-off is adamantine).  Large open areas are best, hallways >2 wide.  Stairways of >2 the reasons being, that half-step diagonal when one dorf passes another is counted and no two dorfs travel at the same speed, especially when hauling.  Hauling jobs stack up and take precedence over everything else so multiple, close, dedicated stockpiles with plenty of wheelbarrows is the way to go.  With bins.  A Borg-like cube fortress is the best layout.  Non-restricted animals should be limited to a few vermin catchers.  Pots trump barrels, by a lot.  Potash is very valuable.  Unimportant areas can be zone restricted.  Then, it’s how many dwarves to you really want?  140 is minimum for MountainHome.  Breaching caverns should be delayed until the fortress is stable with a competent military since there will be more Stuff, more Pathfinding and more importantly, guaranteed sieges from large numbers of Antmen and such (and really, after the first learning fortress nobody plays with Invaders Off, do they?)

Stuff: not all stuff is the same for the counting algorithm.  Trading away pretty much everything not needed is a good idea and something to do with Elves.  Stuff in quantum stockpiles (TradeDepot, slag/garbage pits) doesn’t seem to count as much as stuff lying about or in standard stockpiles.  When a caravan is sighted, it’s an opportunity to scour the surface for single bolts, teeth and parts – also easier on a 2x2 embark.  Teeth and parts can be stockpiled next to a craftsdwarve’s workshop and consumed in useless decorations easier than trading.

This reminds me I have to post Stonesense screenshots of said fort.

[Linkeron]

Can you have multiple stockpiles set to pull from one stockpile now? Last time I tried this, about a year ago, you could only have one stockpile feed into one other stockpile.

[silentdeth]

yes.

[Tierre]

If it is a problem of number of items and their pathfinding - then why not solve it simply? A building warehouse which packs 10 logs into an item - bundle of logs. Also a pack of metal bars and so one. And make reactions to unpack. Can't be done to items with quality - those would change, but absolutely reasonable for other items. That would make an FPS increase. Also you can make 25 bundles or even 100 (too much i think). This shouldhelp with space but with more micromanagement.

[Linkeron]

yes.

Well, with that, would it be efficient to have receiving stockpiles 1x3 with 3 wheelbarrows feed storing stockpiles? Wheelbarrows would be used to stock the receiving piles, and dedicated haulers could then take items off those stockpiles and put them on the main, much larger stockpiles. More jobs get queued, but would certainly cut down on dwarf travel times as the wheelbarrow haulers can move at normal speed, and stocking haulers would only have to travel short distances.

[fricy]

Well, with that, would it be efficient to have receiving stockpiles 1x3 with 3 wheelbarrows feed storing stockpiles? Wheelbarrows would be used to stock the receiving piles, and dedicated haulers could then take items off those stockpiles and put them on the main, much larger stockpiles. More jobs get queued, but would certainly cut down on dwarf travel times as the wheelbarrow haulers can move at normal speed, and stocking haulers would only have to travel short distances.

One minor change: use 2x3 or 3x3 feeder stockpiles, 3 tiles are used for storing the wheelbarrows, so you need empty tiles to generate a hauling job. But yes, the idea is sound, especially for feeding minecarts.

[Linkeron]

Really? I've seen Dwarves just leave the wheelbarrows in the same tile they dropped the load off in.