A shove vole farm optimization request

[MinhPham]

So ... i made this design a while ago and didn't have time to test it in survive mode yet (didn't have time to play, dad), but deep down i know it should work as expected :

• Rectangular!
• The design's goal is to eliminate critter relocation errands / farm supply errands (put egg into the incubator).
• Minimize dupes labor required, after voles reproduce it will be sent into drown chamber.
• Maximize meat/vole rate. This design support 5 voles with 15 cycles lifespan, that's 3.3kg or 5300 kcal/cycle for only 2 shove voles around. The baby inside incubator doesn't cost much cpu ticks.
• The vole has a lifespan of 15 cycles : 4 egg cycles + 5 baby cycles + 6 adult cycles, after that we kill the adult for meat. The idea is to have the baby vole inside incubator instead of inside a stable, then we should have 3/5 of it's lifespan inside the incubator, 2/5 lifespan inside the stable.

 

The problem is, the whole block is quite ... big and i still haven't find a way to trim it down while maintain it's properties ( the middle lane can be 1 tile high but that's it. I tried to avoid using buffer/filter gates for the incubator because it isn't accurate on 2x/3x/10x speed ...

Edit : add save fileWuhan ShoveVole2.sav

Edit 2 : add an image

[Hjoyn]

From what I know, this is almost certainly a more performance optimized build than what you have. Dupe pathing is far more costly than critters, so it's worthwhile to cut out incubators to save on dupe pathing. Shove vole pathing is also minimized, so that has less of an effect as well. The savings in dupe pathing should easily outweigh the cost in vole pathing. 

[Occam Blazer]

Nice build. I don't know of a way to determine which specific critter laid an egg unless they get their own apartment like you have here.

Have you seen this build?

 

[MinhPham]

8 hours ago, Hjoyn said:

From what I know, this is almost certainly a more performance optimized build than what you have. Dupe pathing is far more costly than critters, so it's worthwhile to cut out incubators to save on dupe pathing. Shove vole pathing is also minimized, so that has less of an effect as well. The savings in dupe pathing should easily outweigh the cost in vole pathing. 

• Everyone is capable of building a working shove vole farm (at least the ones with 2+ working brain cells).
• Without killing mechanism the voles get to live 12+ starving cycles.
• Without leaving baby vole inside incubator, the voles spend their childhood outside.
• Sum these up you have 4x voles around to do "path finding", also decor calculation and other stuffs, you may not believe me but they might check whether they can eat the regolith/iron ore chunk on the other side of the map.
• Scale up problem : yes you don't care about optimizing voles ranch until you had to feed a lot of dupes while trying to minimize it's impact. Why a lot of dupes ? well, because there's nothing else to play and it's pretty much the last personal achievement (400+?)....

1 minute ago, Occam Blazer said:

Nice build. I don't know of a way to determine which specific critter laid an egg unless they get their own apartment like you have here.

Have you seen this build?

 

I did ...  but this build requires manual critter relocation + egg supply into incubators, this build also didn't suggest the amount of incubator required for the build in both case : leaving the baby vole inside the incubator, or have the baby inside the stable (which make the stable useless during it's 5 baby cycles).

I also tried in on survival mode, 250 voles ... after dozens of cycles their polulation declined because automation messed up, and dupes throwing voles everywhere

[Occam Blazer]

10 hours ago, MinhPham said:

I also tried in on survival mode, 250 voles ... after dozens of cycles their polulation declined because automation messed up, and dupes throwing voles everywhere

I did this in survival too (-: but modified the design to sweep eggs. Still required dupes to relocate critters. Don't have screens of my condo setup, but I remember I built 20 for 16 dupes. Four incubators was enough from what I remember.

[Hjoyn]

15 hours ago, MinhPham said:

• Everyone is capable of building a working shove vole farm (at least the ones with 2+ working brain cells).
• Without killing mechanism the voles get to live 12+ starving cycles.
• Without leaving baby vole inside incubator, the voles spend their childhood outside.
• Sum these up you have 4x voles around to do "path finding", also decor calculation and other stuffs, you may not believe me but they might check whether they can eat the regolith/iron ore chunk on the other side of the map.
• Scale up problem : yes you don't care about optimizing voles ranch until you had to feed a lot of dupes while trying to minimize it's impact. Why a lot of dupes ? well, because there's nothing else to play and it's pretty much the last personal achievement (400+?)....

Shove voles do not give decor, and with sealed rooms they won't be pathfinding across the map. Keep in mind that all creatures pathfind to all possible locations all the time, which is why dupes are so bad. If a dupe has access to an incubator, nevermind enough incubators to store all the eggs and baby voles, it will be pathfinding to every single one every frame. That is why it's better to just let the voles free in a small box, their highly limited pathfinding far outweighs the pathfinding of dupes if you were to use incubators. Even if it were 4x more vole pathfinding, it would still be worth it. As for scaling up, one dupe with one grooming station supports around 20 dupes worth of barbeque, in that one room. The incubators needed for the same amount of voles would use vastly more space and thus create much more complex pathfinding. 

[MinhPham]

1 hour ago, Hjoyn said:

The incubators needed for the same amount of voles would use vastly more space and thus create much more complex pathfinding. 

Don't give every dupes access to these rooms ... then we can have less path finding complexity for dupes, and less critter to "calculate".

But the way i understand about path finding is a lot difference :

• Only critters that are not moving will do path finding.
• Critters that are moving don't do path finding.
• The less navigation path critters have, the more path finding they will do, since they had to stop more. The vole with only 2 tiles to move will need more path finding than the vole with 20 tiles to move.
• Dupes only do path finding once they take an errand, the path finding function will generate an array of paths and they just follow the path.
• Idle dupes might wander the entire map, and have a chance to do path finding to another location, or just stand still.
• Idle dupes inside a closed base with atmosuit dock entrances can only wander inside the base, but they will mostly wait at the dock.

One more thing ... shove voles have a decor value of 0, but it doesn't mean they don't have decor value to calculate. The same goes for critters trapped in 1 tiles, they can't move doesn't mean they don't do the path finding

[Hjoyn]

1 hour ago, MinhPham said:

Don't give every dupes access to these rooms ... then we can have less path finding complexity for dupes, and less critter to "calculate".

But the way i understand about path finding is a lot difference :

• Only critters that are not moving will do path finding.
• Critters that are moving don't do path finding.
• The less navigation path critters have, the more path finding they will do, since they had to stop more. The vole with only 2 tiles to move will need more path finding than the vole with 20 tiles to move.
• Dupes only do path finding once they take an errand, the path finding function will generate an array of paths and they just follow the path.
• Idle dupes might wander the entire map, and have a chance to do path finding to another location, or just stand still.
• Idle dupes inside a closed base with atmosuit dock entrances can only wander inside the base, but they will mostly wait at the dock.

One more thing ... shove voles have a decor value of 0, but it doesn't mean they don't have decor value to calculate. The same goes for critters trapped in 1 tiles, they can't move doesn't mean they don't do the path finding

I'm assuming only one dupe has access to the incubators, which is still too many. Dupes have to pathfind to all errands at all times, in order to decide which they will take, even while idle. That's why dupe pathfinding is such a massive hit to performance, especially with lots of debris. This is why reducing dupe pathfinding is such a high priority, even at the cost of some critter pathfinding. 

[MinhPham]

8 hours ago, Hjoyn said:

I'm assuming only one dupe has access to the incubators, which is still too many. Dupes have to pathfind to all errands at all times, in order to decide which they will take, even while idle. That's why dupe pathfinding is such a massive hit to performance, especially with lots of debris. This is why reducing dupe pathfinding is such a high priority, even at the cost of some critter pathfinding. 

Dude you obsession with path finding is just too much ... Dupes don't do path finding every frame, only when they take an errand / after they reach the target of the delivery errands. There're one exception that every dupes do pathfinding at once : nagivation path changed, such as when you dig / build / deconstruct a tile.

[Hjoyn]

On 6/4/2021 at 3:59 AM, MinhPham said:

Dude you obsession with path finding is just too much ... Dupes don't do path finding every frame, only when they take an errand / after they reach the target of the delivery errands. There're one exception that every dupes do pathfinding at once : nagivation path changed, such as when you dig / build / deconstruct a tile.

Pathfinding is the primary cpu load at almost all times, and especially for shove vole ranches. As such, if you wish to optimize a shove vole ranch, then optimizing pathfinding is of paramount importance. And while it does indeed seem dumb at times, the pathfinding is stateless and thus works that way. Feel free to verify with the game code if you do not wish to trust me.