(SPOILERS!) Space Exploration 0.6.101 / 1,358 spm / mining 867 / 11.7k hours

[spiral_power]

The plan is to migrate to 0.6.

1. Assembling 0.6 production chains using helmod ( done )
2. Reserve SP for added research
3. Adjustment of some station interfaces to facilitate migration to 0.6
4. Set up resource mines just in case the supply-demand balance changes drastically (iron, copper, coal, bery, holm, irid, methane)
5. Reduce the 0.5 factory from 1,360 spm to 640 spm

6. Stop the 0.5 factory and switches to 0.6
7. Do some research added
8. Install a new 0.6 facilities and gradually bring them on line (expect outages of several days).
9. Optimization and aim for higher spm

[ptx0]

I record the current resource balance as it will be updated to 0.6 in the next couple of months.
At 0.5, there is an excess of iron, copper, stone and coal.

The condition is to perform continuous infinite mining research, automatic collection of arcospheres and, on rare occasions, creation of modules.
Methane was not used because it is finite and under poor conditions.

as I understand it, core mining is very nerfed in .6

and productivity modules are now allowed in many more recipes / miners in space.

and space ores change up a bit

[spiral_power]

as I understand it, core mining is very nerfed in .6

and productivity modules are now allowed in many more recipes / miners in space.

and space ores change up a bit

Yes, it is. I have been following the closed test on discord since 0.6.1 and have verified it myself, so I am aware of the changes. However, my mining efficiency is +70x.
Basically, it is hard to migrate a factory optimized for 0.5 to 0.6. Since a non-stop migration would be very tedious, I decided to allow for several days of outages.

[mrvn]

Finally did a more "scientific" approach for comparing the ups cost of warehouses.

hmmm, it's not that bad considering the benefits that you could get from using warehouses (ex. more inserters and machines can be fitted beside it).

this isn't a realistic test. the items that sleep or stay awake depend on how full the belts are. partially-filled belts, partially-filled warehouses. these are realistic to have. and they change the performance substantially.

you also won't get realistic results with everything in nearby chunks, without a whole lot of other entity types around for them to work from.

as it is, what you've executed looks like a micro-benchmark. it's helpful, but only to show that if you keep scaling the warehouse build up in a real map, you're going to have a bad time.

you can always just reduce the slot count of the containers.

If you have longer belts then it helps a lot to compact them so that there are no holes. Belts are optimized to handle runs of items, you basically just pay for the number of gaps in the belt.

As for warehouses taking so much time, what the game needs are warehouses with very few slots. For unloading iron ore for smelting one doesn't need 1000 slots. 40-80 slots should be perfectly fine. Or even less.

[spiral_power]

If you have longer belts then it helps a lot to compact them so that there are no holes. Belts are optimized to handle runs of items, you basically just pay for the number of gaps in the belt.

As for warehouses taking so much time, what the game needs are warehouses with very few slots. For unloading iron ore for smelting one doesn't need 1000 slots. 40-80 slots should be perfectly fine. Or even less.

First, I will not cheat, e.g., add mods or change settings to make UPS better. I have a policy of using only SE recommended mods. For example, i.e. I do not use loaders. The belts are optimized ( I thought they were uniformly highly efficient up to 200 ? ) I know that the belt in/out is done with inserters, which I consider somewhat costly.

It is not easy to determine which way is best for an entire factory, not just a small test tool. This opinion has been stated previously.

[spiral_power]

By the way, if the item has low flow rates, we don't need 8 6x6 chests for 8 wagons; 1 6x6 chest and 8 stations would be better for the UPS, I do that sort of thing to some extent.

1.png (337.43 KiB) Viewed 4026 times

[mrvn]

If you have longer belts then it helps a lot to compact them so that there are no holes. Belts are optimized to handle runs of items, you basically just pay for the number of gaps in the belt.

As for warehouses taking so much time, what the game needs are warehouses with very few slots. For unloading iron ore for smelting one doesn't need 1000 slots. 40-80 slots should be perfectly fine. Or even less.

First, I will not cheat, e.g., add mods or change settings to make UPS better. I have a policy of using only SE recommended mods. For example, i.e. I do not use loaders. The belts are optimized ( I thought they were uniformly highly efficient up to 200 ? ) I know that the belt in/out is done with inserters, which I consider somewhat costly.

It is not easy to determine which way is best for an entire factory, not just a small test tool. This opinion has been stated previously.

That was more a comment on the warehouse mods themself. Often (for train stations especially) you don't need the large storage but you want the large physical size. A single chest for each wagon.

As for belts just using inserters isn't always enough as gaps can move past the inserter while it swings and you can get some odd oscilations between multiple inserters leaving gaps. The last assembler in a row should unload to a transport belt that side loads to the main belt so it always has a few items waiting to merge into any gap that comes along. Obviously that only works if you produce a full belt (or better just a little above).

There is a debug setting to show the gaps in belts. Try turning that on and checking your belts.

[spiral_power]

That was more a comment on the warehouse mods themself. Often (for train stations especially) you don't need the large storage but you want the large physical size. A single chest for each wagon.

I understand what you are saying. My point is that it is not my job to control the number of slots in a chest. That is just a game condition given to me.

There is a debug setting to show the gaps in belts. Try turning that on and checking your belts.

This is it. Yes, when putting items in a belt and taking items out of a belt, inserters do not finish the process instantaneously as they do for chests. It takes a bit of time. That's what I meant.

1.png (2.36 MiB) Viewed 3962 times

[mrvn]

That belt at the top going to the right and up looks bad. But that might be just temporary.

If you have many such belts between parts of your factory then that is worth optimizing.

[spiral_power]

What would that optimization be regarding, it's not UPS, is it? There may seem to be a gap at the moment, but during the cycle of the system, supply and demand are stable and there is no excess or shortage of items.
I don't know about others, but I always aim for a system that works consistently and forever.
I am not claiming that I am the one who temporarily achieved 1,360 spm, just that I cut out the convenient part of it.

[Stringweasel]

If you have longer belts then it helps a lot to compact them so that there are no holes. Belts are optimized to handle runs of items, you basically just pay for the number of gaps in the belt.

Apparently the gaps on belts are negligible for UPS. I thought the same, but some Factorio's UPS optimization experts told me it's not the case, or at least the point is still up for debate. There might be more segments to update in the belt, but the impact is tiny, because it's so fast. It's hard test this of course. And we don't have access to Factorio's code either.

Could be that adding more machines to keep the belt compressed will have a larger impact on UPS anyway.

So IMO it's definitely not worth it to optimize something that's up for debate, and hard to optimize.

Anyway, epic looking base! Will you be swopping your train-transporting-spaceships with space elevators in 0.6?

[mrvn]

If you have longer belts then it helps a lot to compact them so that there are no holes. Belts are optimized to handle runs of items, you basically just pay for the number of gaps in the belt.

Apparently the gaps on belts are negligible for UPS. I thought the same, but some Factorio's UPS optimization experts told me it's not the case, or at least the point is still up for debate. There might be more segments to update in the belt, but the impact is tiny, because it's so fast. It's hard test this of course. And we don't have access to Factorio's code either.

Could be that adding more machines to keep the belt compressed will have a larger impact on UPS anyway.

So IMO it's definitely not worth it to optimize something that's up for debate, and hard to optimize.

Anyway, epic looking base! Will you be swopping your train-transporting-spaceships with space elevators in 0.6?

Maybe it doesn't even improve UPS. But there is a second factor to consider: A belt with gaps carries less items than possible.

So don't go adding more assemblers to your units. But if you already produce (more than) enough and have gaps in the belt then adding a little side loading to compact the belt will increase the number of items transported. If it's just a matter of placing the last inserter differently then I would definetly go for it.

[ptx0]

we're going pretty off-topic for this thread. but consider that inserters trying to dump item onto a compressed, moving, belt, will stay awake until a position is found. the backed-up belt doesn't wake inserters that are waiting for a spot, if it doesn't move. and a non-compressed belt doesn't wake inserters, because they're asleep waiting for assembler or chest to produce an output.

[mrvn]

Urgs, now we are getting into secondary and tertiary effects. Yeah, this is going really off topic.

[spiral_power]

So IMO it's definitely not worth it to optimize something that's up for debate, and hard to optimize.

Yes, I totally agree. I don't want to optimize something that I don't know will be obviously better.
It is in these obvious cases that optimization is worthwhile.
https://forums.factorio.com/viewtopic.php?p=562792#p562792

Anyway, epic looking base! Will you be swopping your train-transporting-spaceships with space elevators in 0.6?

Thank you.
Because of the position (coordinate) constraints of space elevators, the existing factory would need to be significantly reworked in order to implement it. I cannot introduce it right away, but I intend to do so eventually, as UPS should be clearly advantageous.

Maybe it doesn't even improve UPS. But there is a second factor to consider: A belt with gaps carries less items than possible.

So don't go adding more assemblers to your units. But if you already produce (more than) enough and have gaps in the belt then adding a little side loading to compact the belt will increase the number of items transported. If it's just a matter of placing the last inserter differently then I would definetly go for it.

I have already done everything possible to optimize the number of assembly machines and to maximize the use of belts.


I have no problem with a little off-topic. It is a learning experience for me and I will continue to post my images.

[blazespinnaker]

Thanks for sharing. I went through all the pics. Very incredible and much respect. Best I ever got to was around 800 spm or so, and not very sustained, though I did get it to a fairly stable point with no deadlocks anywhere running overnight. My arcosphere setup was pretty basic though, I just put all the logistic stuff close together and put limits on logistic requests if supply was over X. Dead simple, decent throughput.

TBH tho, I sort of quit when I got the arcolink stuff. For some reason it just spoiled the mod for me.

Fortunately, some fairly dramatic changes in the latest updates, so a good reason to restart Good challenges, though I very much dislike the lack of PM modules in casting, and removal from ignots. No PM in space made sense, but seems a bit retentive to limit them planetside. Everytime I look at those factories I just grimace, knowing the best I can do is add speed, which seems rather artificial and would just be doing it for UPS, if any. Hopefully he'll revisit that choice.

[blazespinnaker]

One question about the matter fusion .. I remember looking into it, but the issue was really only naq and vita, everything else was relatively infinite already and I would only have done it for kicks. There was some issue with raw resources as well, can't remember, water? Or stone? Couldn't find an asteroid field that had the right mixture, anyways.

Of course, everything is pretty much different now. Feels a bit unbalanced, tbh, but I really don't mind, just makes for interesting challenges. Sort of like the randomizer mod for recipies. Like I said, my only complaint is the PM thing.

[spiral_power]

Thanks for sharing. I went through all the pics. Very incredible and much respect. Best I ever got to was around 800 spm or so, and not very sustained, though I did get it to a fairly stable point with no deadlocks anywhere running overnight. My arcosphere setup was pretty basic though, I just put all the logistic stuff close together and put limits on logistic requests if supply was over X. Dead simple, decent throughput.

Thank you.
These are if they are stable enough.
I think 800 spm is pretty high. Is the subject of your research mining? Other than that? The mining research I am doing is a little easier since there are only two types of T4.
I am also interested in how many arcospheres are used in the balancers.

One question about the matter fusion .. I remember looking into it, but the issue was really only naq and vita, everything else was relatively infinite already and I would only have done it for kicks. There was some issue with raw resources as well, can't remember, water? Or stone? Couldn't find an asteroid field that had the right mixture, anyways.

Matter fusion is very slowly, power hungry and consumes a lot of chemical gel (i.e. crude), so I don't think it is necessary for normal game clearing. I think it is an extra feature after clearing the game.
Ore generation by matter fusion is very slow. Basically, it is quicker to dig out the ore itself that is needed. Even if they need to be transported from other locations.

However, matter fusion may not be advantageous with UPS. A good UPS is about the third most important thing to me. More important is the stability to keep it running forever, maintainability and it's SE unique features.

Like I said, my only complaint is the PM thing.

This is a 0.6 recipe, but did you know that there are high performance facilities available for space use? It is much better than that new little one.

2.png (1.24 MiB) Viewed 3729 times

[blazespinnaker]

Yes, it was mining research. Arcospheres, I think was around 80 or so.

I'm not entirely clear on what you mean by above

"This is a 0.6 recipe, but did you know that there are high performance facilities available for space use? It is much better than that new little one."

The 'little one' is quite compact, but without prod modules it is very static. I could waste more speed modules on it, but that just feels wrong. Sending molten to space for casting might be doable with a very large spaceship, but that's hardly UPS efficient and problematic in many regards. Maybe the PM module limitation in space has been removed somehow, I haven't checked, but that would be weird. The whole point of the PM weirdness was to build planetside and ship to space, which was awkward but an interesting challenge at least.

No PM modules in the new casting machine is pretty silly though. Removing them from the ignot is probably survivable because sending them up the well makes sense anyways.

Maybe there is some magical teleporting pipe thing I haven't bothered discovering yet, but if it's anything like the arcostorage, I hope I don't. I really dislike magical teleportation, totally ruins the game for me.

[spiral_power]

Yes, it was mining research. Arcospheres, I think was around 80 or so.

80?
Perhaps my question did not come across well.
For example, this recipe uses 64 arcospheres.
Overall, it is about 1.5k, as I wrote at the beginning of the thread.

1.png (389.45 KiB) Viewed 3628 times

I'm not entirely clear on what you mean by above

"This is a 0.6 recipe, but did you know that there are high performance facilities available for space use? It is much better than that new little one."

The 'little one' is quite compact, but without prod modules it is very static. I could waste more speed modules on it, but that just feels wrong. Sending molten to space for casting might be doable with a very large spaceship, but that's hardly UPS efficient and problematic in many regards. Maybe the PM module limitation in space has been removed somehow, I haven't checked, but that would be weird. The whole point of the PM weirdness was to build planetside and ship to space, which was awkward but an interesting challenge at least.

No PM modules in the new casting machine is pretty silly though. Removing them from the ignot is probably survivable because sending them up the well makes sense anyways.

Maybe there is some magical teleporting pipe thing I haven't bothered discovering yet, but if it's anything like the arcostorage, I hope I don't. I really dislike magical teleportation, totally ruins the game for me.

While I was working on it, the sentence seemed a little strange.
I may have been brief in my explanation.

little one pattern.

2.png (1.16 MiB) Viewed 3628 times

Using a spaceship floor, facilities for space can be set up on the ground, although the recipe is limited. The number of entities is smaller than the little ones, so the problem of PMs not being available may be somewhat better. This is the preferred method from a UPS perspective.

3.png (3.03 MiB) Viewed 3628 times