Automatic Piler

[Sine]

ok it's more than likely that i'm the only one but... can anyone make an exemple where i can use this thing? becouse i'm sorry to say that this machine dosen't make any sense to me

[margalus]

Anywhere your belts are running at 100% capacity, but you need more. Easy to way to double capacity or more if I understand it correctly. I haven't actually tried it yet though.

[josmith7]

Yep, I played with it a little. In the best case you can stack cargo up to 4 high, allowing a Mk.III belt to carry 7200/min instead of its normal 1800/min. (Of course that means you need 4 full Mk.III feeder belts running into your stackers; so it's only useful when you want a fairly long very-high capacity belt instead of just running those 4 belts the whole way)

That also makes sorters more efficient because they're now moving up to 4x as much.
It's a bit tricky to arrange the stacking appropriately thought.

One example where it might be useful is producing Casamir crystals. A normal Mk.III belt can only carry enough deuterium for less than 7 Mk.III Assemblers. But if you did some stacking you could run several belts into 1 and have that 1 supply maybe 25 assemblers.

Another example would be with stacked input belts you could probable load an ILS by belt fast enough to keep up with the rate drones or vessels can empty it. In that case it'd effectively let you jam up to the equivalent of 48 normal belts into its 12 belt slots.


But basically any time a belt is a bottle neck you could stack the material to boost the belts throughput.

(Note however that as of this update normal miners, and probably the new advanced one though I haven't tested that, already output up the 4 stack -- so those belts can 'natively' handle up to 7200/min mined material)

[Spynal Toad]

After running some tests, I've come to some conclusions:

1 - the Piler DOES NOT wait for extra items to come along. if a singular item comes and is not swiftly followed by another of its kind, the single item will be shepherded through.

2 - the arrows that light up correspond to the tier of the belt which is its "output" - HOWEVER, i also observed that if both input and output belts are T3 (Blue), the arrow appears as GREEN, which is T2. which leads me to;

3 - the Piler seems to be expecting to deal with a fully stuffed belt of a higher tier, so it can output a half filled belt of stacks - these stacks are VISIBLY TALLER, so you can tell them apart.

4A - as per the guide, the piler takes two single items (in rapid succession as per point 1 - the piler waits for no one!), and converts them into a single "stack" of two

4B - however, also as per the guide, these double stacks can themselves be stacked, to make a stack of four. but again, it is expecting a lower tier of belt on its output, so a fully stuffed T3 belt's output can be stacked once into a T2 belt carrying stacks of two, and again into a T1 belt carrying stacks of four - the hard upper limit of stacking.

5 - BUT WAIT - if you reverse the input/output (tall side in, short side out) this entire process can be reversed! you can take that T1 belt full of Quads and turn it back into a fully stuffed T3 if you desire it!

6 - (unconfirmed) it seems that sorter arms treat these stacks as a single item, but i'm not 100% sure. plucking them from the belt by hand confirms the stacking process works, however.

[Spynal Toad]

this leaves me with a question: if a recipe calls for two of something, and the sorter arm grabs a full stack of four, but the receiving machine can only take two more of that item, what happens? does it reject the whole stack, or does it perhaps take part of the stack

[Spynal Toad]

Never mind, it seems my question already had an answer, it will simply take the stack and "over-stuff" itself, because by default, a machine will not "fill" itself with needed material, so there is room for such an overdraw in regards to taking resources from a belt!

[KaiserTom]

Yeah, so it's literally a cargo mux/demux. You can get some crazy throughput on a bus now.

How does this play with sprayers? Does a 4 stack eat 4 sprays or only 1 spray?

[kris44dad]

Originally posted by Sine:

ok it's more than likely that i'm the only one but... can anyone make an exemple where i can use this thing? becouse i'm sorry to say that this machine dosen't make any sense to me

If Steam will take a link, I just made Casimir crystal without all the H2 lines.
https://www.dysonsphereblueprints.com/blueprints/casimir-crystals-from-ogc-with-pilers-450-min

[Dirkels]

It can be help full for those recipes that need 10 or 12 items per cycle, now you can have more of them on a single belt. Of course you need 4 input lanes and 6 pilers to create 1 fully loaded quad stacked belt.

And now I think of it could it improve the fractionater? To the test bench....

[!?!]

Do those things not kill performance with up to 4 times as much stuff on the belt and irregularities?

Sounds like it isn't a "4 stack" item with a custom model but 4 individual items drawn on top of each other.
And then it also doesn't even wait until it has 4 items and always outputs a stack of 4, it just farts them out whenever so a backed up belt is fragmented into little clusters of differently stacked items which seems like optimization poison?

Well at least the reason factorio runs so well is because full belts of the same stuff gets consolidated internally into chunks, stacking of this kind would kill that.

[kris44dad]

I think the biggest issue with the piler, is finding the best belt layout to fully use it's capabilities. For example, it doubles a line, but you need to inject another line after the piler to "fill up" the line again. (stack of 2 on the belt, then one, at that point...) If you then use another piler, presumably that results in a stack of three? Inject again after that one with a final piler to get a stack of 4? It MAY have been a little easier to comprehend if it had three ports, instead of 2. Particularly, if the two ports had the same spacing as the ports into/from the towers.

When I first saw it, I assumed it would take two belts that were placed one over the other, and integrate them into a two stack. It is useful, but if a 2, then 1, on a belt doesn't integrate into a 3, I think you have to use 2 pairs to get a 4 stack. I'm waiting to see that blueprints people share. I'm sure they will think of something I haven't thought of yet.

[Dirkels]

Tested out a fractionator like of 6 fractionators in a row, When fed by a single height blue belt I get about 100 deuterium/min out of it, If I feed it with a quad stacked line and also have 2 pilers (stackers would be a better word) on de return line from end to start I get around 400 deuterium/min. And no need to feed it from multiple input lines once up to speed the insert speed into fractionator loop is so low 2 pilers in a single line is enough to get stacks of 4 high.

[Dirkels]

Originally posted by !?!:

Do those things not kill performance with up to 4 times as much stuff on the belt and irregularities?

Sounds like it isn't a "4 stack" item with a custom model but 4 individual items drawn on top of each other.

Looking at the fluid effect leads me to belief stacks are single objects, just a higher and with separation lines that show how many are in the stack.

[Dirkels]

Also now I see fractionators work significantly better when fed with quad stack hydrogen I feel would like to see an option for filters so you can filter out stacks of a certain height or even better a height and the option <, = or >, Default >0 for current behavior.

[KaiserTom]

Originally posted by !?!:

Do those things not kill performance with up to 4 times as much stuff on the belt and irregularities?

Sounds like it isn't a "4 stack" item with a custom model but 4 individual items drawn on top of each other.
And then it also doesn't even wait until it has 4 items and always outputs a stack of 4, it just farts them out whenever so a backed up belt is fragmented into little clusters of differently stacked items which seems like optimization poison?

Well at least the reason factorio runs so well is because full belts of the same stuff gets consolidated internally into chunks, stacking of this kind would kill that.

The 4 stacks are not 3 more additional objects. It's the same cargo object just with an additional variable that says how much cargo that object is worth, such as 4. Any performance hit from it is very marginal, it's basic integer math. It's just not worth worrying about when other sections of the code take up 99% of the processing time. And a single 4 stack blue belt is going to be much more performant than 4 separate 1 stack belts. Literally a reduction in objects by 1/4th

And they've also optimized those slow parts further to get an overall hefty performance improvement from the update.

I actually have more concern over the performance impact of sprayers, but it also seems like just an attribute being applied to the object. Rather simply

[Xilo The Odd]

new question, does a stack of 4 product only use 1 spray from a proliferator?