[Vanilla|0.14|US] SUG Dedicated Server

[pedter]

- Updated factories to assembler 3s
- There were problems with green circuits not getting enough copper (which only got worse after I upgraded the assemblers) and with red circuits not getting enough copper wire
- The most belt bandwidth is needed for the wires
- I also upgraded the efficiency modules in the furnaces, refineries and chem plants to be level 2
- Added some beacons with level 3 efficiency modules to the largest offending oil field on the map
- Set up another copper source ("Copper 4"), but didn't assign it a train as there are currently only two train stops

- Factory was designed to launch a rocket every 45-50 minutes (10/11/11 rocket component assembler 2's)
- Wasn't designed to go any faster or for any other loads on the system
- Assembler 3's and other loads are now overloading everything: belts, smelters, miners, pumpjacks, and probably trains
- Eff.2 are a waste of resources (eff.3 are even worse); build more solar instead
- Beacons burn more power than they save when fitted with eff. modules unless you're hitting a huge number of offenders; build more solar instead
- 62MW of defense lasers exist for a reason; ignore pollution unless you've nothing better to use than eff.1 (mostly assemblers)
- Cycle a train between an existing stop and the new one to share the load (ex: home -> copper 1 -> home -> copper 4)

There are zid/ziid/zcd/zad stops that someone was apparently working on - I'm kind of curious what those were going to be.

Factory using prod.3 with speed3 beacons. Rocket every 17 minutes for 30% of the resources that the existing factory uses. Will replace the current silo.
Stops are named as such so they don't clutter the top of the train stop list.

Those stops were added by pedter - he was building the dropoff for the rocket silo.... Not sure where it went!

It's in my pockets.

[NotABiter]

Eff.2 are a waste of resources (eff.3 are even worse); build more solar instead

I did not add them to save power, but to reduce pollution. (I did mention the -80% power bonus, but you have to realize power is one of the components of pollution - reducing power directly reduces pollution by the same amount.)

Beacons burn more power than they save when fitted with eff. modules unless you're hitting a huge number of offenders; build more solar instead

I didn't build them to save power, I built them to reduce pollution - as I stated in my post. (Note that I said "biggest polluters", not "biggest energy hogs".)

Assembler 3's and other loads are now overloading everything

I don't really think of it as "overloading". Switching to assembler 3s, along with decreasing power and pollution, increases throughput, and those are all good things. That increased throughput potentially changes where a factory's bottleneck is, just like switching to a faster belt increases that belt's throughput, potentially making it not be the bottleneck anymore. (But as stated, that didn't even happen here - the bottleneck was already the single copper plates belt trying to feed 15 copper wire factories.) Increasing the available throughput of something can't "overload" something else unless the thing you just increased the available throughput of was itself already "overloaded". (If it wasn't, then increasing its available throughput would not increase its demand at all.)

62MW of defense lasers exist for a reason; ignore pollution unless you've nothing better to use than eff.1

I bet the power companies in the US wish they could just use lasers instead of scrubbers and more expensive low-sulfur coal. Telling me to "ignore pollution" is probably like telling you to "ignore how many resources I spend". (Throughput, pollution and productivity are the main things I like to improve. I also tend to play "the long game", spending some resources now to provide factory improvements that continue to provide benefits into the future.)

Cycle a train between an existing stop and the new one to share the load

You could do that. I chose not to (and was thinking it would be brought online when one of the others was exhausted or when more train drop-offs were added). It depends on how you prefer to manage the consumption of the copper fields. But from what I saw that scheme wouldn't improve throughput of the whole system as it would still be bottlenecked on the factory (and after that the train). Copper ore wasn't an issue, except for the couple of times I managed to block the trains. And if it were an issue, multiplexing one train probably wouldn't help - you'd need more trains. My real reason for getting the copper field ready to go was that I was aware of the fact that I sucked up a bit of copper making modules and I didn't want to make a lot of extra work for whoever else played the map after me so I got more copper ready to go.

Factory using prod.3 with speed3 beacons.

Oh my. Will there be hazmat suits available? That's going to be a lot of pollution.
Except for pump jacks, I generally don't see the point of going crazy with speed modules - just make your factory modular and then use construction bots to expand capacity at will. You can then achieve the same throughput* using less power, less pollution, and probably a similar number of resources spent (using speed modules saves needing more productivity modules for more instances, but costs in terms of speed modules/beacons/power since you only get 1/2 effect from beacon modules). Of course, when I use productivity modules (which increase pollution) I also use beacons with efficiency modules to get the pollution back down (and maybe add the odd speed module if I'm already hitting -80% power and have some beacon capacity to spare).

That said, speed modules aren't as bad as they might seem on a pollution-per-item basis because even if it increases power (and therefore pollution) by 70%, the 50% speed bonus means less time spent working (generating pollution) for each item, so the speed partially cancels out the pollution increase. But still, beacon slots used for speed modules are beacon slots not used for efficiency modules. (You can actually get pollution to less than the normal -80% power level by using a combination of efficiency and speed modules, maintaining the -80% power bonus while cranking the speed up a bit.)

*: Actually you may get much better throughput by replication rather than speed modules, because speed modules don't improve the throughput of belts or inserters whereas replication can. With replication, sticking to the cheapest transport belt can even be a viable option.

Does anyone know how long this map will be allowed to continue running? (When he updates to 12.17 will the map be upgraded or reset?)

Maybe I won't do it here since we seem to be at cross-purposes, but I think I might be inspired to try playing a more resource rich (relative to biters anyways) map in SP and make a giganto/modular high-throughput factory.

[pedter]

Oh my. Will there be hazmat suits available? That's going to be a lot of pollution.
You can then achieve the same throughput* using less power, less pollution, and probably a similar number of resources spent
I also use beacons with efficiency modules to get the pollution back down.
But still, beacon slots used for speed modules are beacon slots not used for efficiency modules.

This table shows that speed modules are, quite surprisingly, superior to efficiency modules at reducing power and pollution when boosting an assembler full of productivity modules. Try it: add a half eff.3 beacon to any line in that table and you'll find that putting the next half speed3 beacon in it's place is actually more beneficial. The faster we go, the less power we use! Well, sorta. There's more to it than that.

We care little for power and pollution per second - we care instead about power and pollution per item. If we can double throughput while cutting power and pollution of each item in half we would clearly make the switch yet our power and pollution per second have not changed. In fact, building a more complete table comparing half-modules used in beacons, we do not reach a break-point on a per-item basis until there are already 6 beacons (12x speed3) boosting. At that point, eff.3 modules in any additional beacons barely (and I mean just barely; I had to go out to 4 decimal places) creep out ahead in terms of power and pollution saved.

But, of course, all of that power and pollution talk ignores the real reason we use speed3 in beacons: the factory with prod.3/speed3 (instead of prod.3/eff.3) is actually the factory that's going to use less power, make less pollution, and cost less to build. As a bonus, it requires zero logistics robots because it's so small and zero buffers other than ore and rocket parts because everything else moves so quickly. Compared instead to a factory that skips both prod.3 and speed3, it uses a substantially smaller resource throughput to keep fed.

We've discovered that the current factory is not really limited by what it can produce; everything we ourselves use is buffered and we'd have a difficult time breaking those buffers. That's what we built it for: to keep up with us. It's instead been broken by launching rockets - a rocket an hour is more than fast enough for most of us but we'll only launch a couple before we drain the resource fields again. It's instead been limited all too often by the rate at which we can secure (and have the patience to secure) additional resource deposits while building rockets. Stretching those resources 3-4x as long is incredibly powerful.

Throughput, pollution and productivity are the main things I like to improve.

I used to do that exclusively; then I learned how to actually use productivity modules. I'd never used them before because they looked awful for power and pollution - and they are, at least on their own. I discovered that they fit quite well within my efficiency outlook when used with speed modules, however. Now? I still do that exclusively. I've just discovered that others found a way to make it even better so I adopted it, though I still haven't found the time to deploy it.

Tetyron has put in many hours expanding the walls; it's been vitally important to the other half of combating pollution. The walled-in base is expansive enough that the total pollution is nearly, if not fully, absorbed by the terrain before it reaches the walls. When and where it isn't expansive enough, we expand again. Our standard has always been 2x eff.1 in assembler 2's, furnaces, chem plants, refineries, and pumpjacks to make it easier to absorb the pollution. Assembler 3's were an extension to that; they receive 3x eff.1 and, only if necessary for throughput, 1x speed1. Pumpjacks posed another problem; they wouldn't stand a chance against the rest of the factory if we weren't boosting them with +100% speed, so we made the switch from eff.1 to speed3.

[NotABiter]

This table shows that speed modules are, quite surprisingly, superior to efficiency modules at reducing power and pollution when boosting an assembler full of productivity modules.

Not exactly. That table shows a bunch of numbers that may or may not have anything to do with the behavior in the game. Not to question your math/facts but...

In fact, building a more complete table comparing half-modules used in beacons, we do not reach a break-point on a per-item basis until there are already 6 beacons (12x speed3) boosting. At that point, eff.3 modules in any additional beacons barely (and I mean just barely; I had to go out to 4 decimal places) creep out ahead in terms of power and pollution saved.

...I have to question your math/facts, because that (having "to go out to 4 decimal places") is not what happens in the game.

Note that efficiency modules and speed modules only affect speed and power, not the "base pollution rate" (unlike productivity modules which do affect it), so their affect on pollution is fully captured by power integrated over time, which is energy. You can measure energy use quite directly in the game. Here's the test jig I just used:

test jig

energy_test.jpg (144.15 KiB) Viewed 9071 times

The beacons are on the main power grid. Only the assembler is powered by the accumulators. The assembler contained 4 productivity module 3s for all of the tests. A test consists of the following steps:
1. fully charge the accumulators with a substation
2. pick up the substation so the accumulators are no longer powered by the main grid
3. drop 100 iron plates into the assembler
4. wait for the assembler to finish
5. click on one of the medium poles to see how much energy was consumed from the accumulators

Results are as follows:

6 beacons:
* 12 efficiency module 3s: 13.2 MJ
* 12 speed module 3s: 11.8 MJ

7 beacons:
* 14 efficiency module 3s: 7.9 MJ
* 14 speed module 3s: 11.2 MJ

8 beacons:
* 16 efficiency module 3s: 2.3 MJ
* 16 speed module 3s: 10.4 MJ

As can be seen above, with enough beacons efficiency modules completely trounce speed modules in terms of total energy used (and therefore in terms of pollution generated). Also, at 16 efficiency module 3s, power use is already pegged at -80%, so just adding more beacons and efficiency modules won't improve things any further, you have to add a mix of efficiency modules and speed modules to see further improvement (as noted in my previous post).

Also, as you pointed out, module 3s are expensive so I often don't use very many of them and mainly stick to 2s. Even with level 2 modules, efficiency modules are still a big win for reducing pollution, but you need a few more beacons to reach that "magical" -80% power usage. That's why, for example, in the last game I finished I had 12 beacons affecting each assembler. (Picture and info are here.)

[pedter]

...until there are already 6 beacons (12x speed3) boosting.

This bit is important. I stated that 6 beacons is the breakpoint; you just proved it. At 6 beacons, your speed3 setup is superior to your eff.3 setup. Beyond that, eff.3 trumps in terms of power and pollution because energy useage drops below 100% per unit time. However, you again ignore the fact that at 6 beacons, your factory is 8.5x as large; at 7 beacons, 9.75x, and at 8 beacons, 11x as large as the speed3 factory. That breaks computers, even in 0.12, when I'm launching a rocket every 56 seconds.

I have to question your math/facts, because that (having "to go out to 4 decimal places") is not what happens in the game.

I gave this one a test myself and I can confirm the 13-module breakpoint - 7 beacons, using either 13x speed3 or 12x speed3 / 1x eff.3. For 13x speed3, listed values are 1837.5kW, 47.25 pollution, and 4.5625 crafting speed. 12x/1x shows 1711.5kW, 44.01 pollution, and 4.25 crafting speed. Taking the ratio, you must go to 4 decimal places - 1.073 doesn't cut it. Going a full 5 instead shows 1.07353x for crafting speed and 1.07362x for power and pollution. 12x/1x just barely creeps ahead of 13x.

However, letting it actually run (and I tried multiple times) is doing weird things. Casting 50x 0.5sec iron gears takes 5.48 seconds with 13x; this burns 10068kJ of power. However, accumulators show 10.8MJ consumed; somehow we've burned extra. It gets even worse with 12x/1x; it takes 5.88 seconds, or 10067kJ, yet the accumulators show 11.5MJ consumed. Testing using 13x eff.3 in the beacons isn't any better; 0.5 crafting speed at 199kW takes 50 seconds, or 9950kJ, yet the accumulators show 10.6MJ used. There's nothing else on the system - no lights, no beacons, no inserters. The extra power consumed varies quite a bit, but isn't a function of how long we spent crafting.

Also, as you pointed out, module 3s are expensive so I often don't use very many of them and mainly stick to 2s.

I'll assume you leave prod.3 modules in - expensive, yes, but you need a limited number of them and you can't just add more prod.2 to get the same resource savings as if you'd used prod.3. Then, in the case of using eff.2 and speed2, you'll see pollution tip over in favor of eff.2 when you exceed 7 beacons. Your factory again ends up nearly an order of magnitude larger, however. Power usage goes up by a similar amount because of all of the beacons.

[NoPantsMcDance]

Updated to 0.12.19. Sorry for the delay on updating been busy doing other things. Enjoy

[NoPantsMcDance]

Updated to 0.12.20 along with new map. Enjoy!

[tetryon]

Thanks for updating. I've been able to log on just fine since the update up until now. Never had a problem before the update.

Now on this, and another server, I'm in a desync loop.

Are you able to check/restart the server?

I wish I knew what caused it!

Thanks for running this

[NoPantsMcDance]

Thanks for updating. I've been able to log on just fine since the update up until now. Never had a problem before the update.

Now on this, and another server, I'm in a desync loop.

Are you able to check/restart the server?

I wish I knew what caused it!

Thanks for running this

Just gave it a restart and checked for any loop. It appears to be fine but not too sure. I was having the same issue earlier but other people were playing fine so might be a bug with the new update. Let me know if it worked for you.

[elahrairah]

I was on for a while yesterday. (Winx64)
Late last night I installed to my mac and tried to connect, and found it doing the desync loop.
Today, on my Windows machine it's also doing the desync loop. THere's a few other people on, so it's not so much the server. I think there's 3 people on successfully, but I can't get in.
Seems 12.20 is full of desyncs.

[GoldenPorkchop80]

So, I happened to try to join just now.

It seems the server is stuck in a desync loop.

Hopefully you can get it fixed? I really like playing on your server!

Thanks!

[NoPantsMcDance]

So, I happened to try to join just now.

It seems the server is stuck in a desync loop.

Hopefully you can get it fixed? I really like playing on your server!

Thanks!

Restarted the server it should be fixed for now. 12.20 seems to be having a ton of desync issues atm sorry for the inconvenience

[GoldenPorkchop80]

Restarted the server it should be fixed for now. 12.20 seems to be having a ton of desync issues atm sorry for the inconvenience

It's ok!

I mean, after all, 12.20 is a stable candidate!

Well, I will keep on playing! Thanks again!

EDIT: 12.20 is now a Official Stable Version! Huzzah!

[elahrairah]

Griefed again. Whole base deconstruction-plannered.
This causes 3,000 bots to go nuts and that causes a DESYNC loop. I can't even get in to begin rebuilding.

Need a rollback to autosave, or a fresh map. 12.20 is not stable.

[NoPantsMcDance]

Griefed again. Whole base deconstruction-plannered.
This causes 3,000 bots to go nuts and that causes a DESYNC loop. I can't even get in to begin rebuilding.

Need a rollback to autosave, or a fresh map. 12.20 is not stable.

Server is back up.

[antanis317]

Was a world reset not planned a few Fridays ago?

[pedter]

Was a world reset not planned a few Fridays ago?

Reset happened 12/5, so about two weeks ago.

[wesleytech]

uhm today i tried to join, but something went wrong... i have a desync report if you need it.

[juani23]

hello i have a problema i cant conect to server it say couldt stablish network communication with server.
someone know why it happen

[NoPantsMcDance]

hello i have a problema i cant conect to server it say couldt stablish network communication with server.
someone know why it happen

I restarted the server so it should work now. I'll add a daily restart to my script so it refreshes often.