UPS Optimization - Red Science (editor)

[Stevetrov]

I remember looking closely at those monolithic 1K SPM blocks from smurphy1 and fallenghostplayer a while back, and they are impressive feats of engineering. But pretty much immediately my mind wanders to, "how would I even begin to provide 65K+ iron ore per minute to all the conveniently placed inputs around the perimeter of this thing, in a real-world map where I don't already have hundreds of levels of mining productivity?" So it made me wonder if adding some constraints around that would make these builds more interesting to a wider audience.

These designs are intended for extreme late game, where everything is optimal for building a high UPS megabase ie you have large patches available from settings or exploring far from the starting point and you have researched a reasonable amount of MP science.

Here is an example of my 1K cell build on assemblystorms server (vanilla with increased but not maxed resource settings)

Reallife_megabase.png (287.46 KiB) Viewed 2094 times

Before you get to this stage builds with ore delivered by train are a good compromise between UPS and practicality.

That's why I designed this build, its 500SPM, takes ore by train and doesn't mind too much if some trains have partial loads. https://www.reddit.com/r/technicalfacto ... belts_and/
I am fairly sure this design could be improved significantly and maybe making improvements to this could be a contest, but building an entire UPS optimised base feels too big for a competition like this.

[Belter]

The one thing that bothered me a bit about these tests was allowing for arbitrary ore patch placement via the editor.

These designs are intended for extreme late game, where everything is optimal for building a high UPS megabase

I'd like to thank you for both posting this, maybe others feel the same way. I'd join Stevertrov of course. Factorio is actually several games in one from the early games to megabases. If you want to scale up further you find yourself tweaking map with high density ore patches and/or turning off enemies as your potato computer cannot run the game w/60 UPS anymore. Or just find military repetitive. That's already not the "original" game, not to mention the waste number of mods...

I take playing in editor mode the final, vanilla stage of this game. This is the end. But everybody is different. I only have a 2.7k spm megabase full of obsolete sub-factories and find UPS optimnization a big fun. But got really tired of brining the resources to the construciton, making a bot network first so they can build etc. Actually take a look on Shapez which is kind of even more stripped down game of Factorio. OK it looked that way until floating shapes came...

some valuable learning still happened [...] I'm imagining a blue science build that requires iron/copper/coal/oil/water patches of just the right size and location to support optimal DI layouts...

Exactly! one the weekend I'll start the green science thread here, hope there will be interest. I've learned a LOT from these threads, maybe you could take a look on the iron smelting one also which started as a "normal" one. Also the way those blueprints / savegames are shared could be motivating for others and seeing the HUGE UPS difference might be useful for some folks who has assumptions but no time/paitence to test them out.

Finding the most efficient ways to buffer and distribute materials delivered by train to feed these DI builds seems like a worthwhile next step.
Maybe the problem space becomes too big that way and isn't fun. Or maybe I'm missing the point here. Anyway, curious to hear people's thoughts.

I'm on the "isn't fun" side. I did myself two way train grid with proper signalling which was a fun. Then realized that my way is the 1 base / pack to separated things. I like to divide the problem (which is - for me - UPS optimization) and ruling out ore distribution is a logical step forward. I'm sure there would be an interest in measuring distributing methods also.

The only thing really bothers me that UPS is scaling differently on different CPUs. This is something we have to live with and accept that 1%+ inconsitecies.

[Tertius]

I take playing in editor mode the final, vanilla stage of this game. This is the end. But everybody is different. I only have a 2.7k spm megabase full of obsolete sub-factories and find UPS optimnization a big fun.

(Only because you as thread starter included it into the discussion:) This is much similar and at the same time very different to what I find fun. I also do almost everything in editor mode with only QoL mods but no vanilla mechanics changing mods. My current project (since a year) is also a 2.5k spm megabase.
However, it's not the ups optimization I find fun - I got to it because of CPU constraints, but not because it is a genuine challenge the game was made for and with.
Now I've got a more powerful PC, the requirement for UPS optimization vanished into thin air. And I don't spend one more thought about it. I realized my fun is with designing factories that use the least ingame resources to produce their goods. Least amount of machines produce maximum amount of output with the least possible space. That's what I'm optimizing, and that's what I find big fun.

I watch all these ups optimization threads and find them very interesting, and I see certainly new concepts I will put to good use. But if I see the builds in this thread, I feel these are abominations with the only intent to stress the game engine as little as possible no matter the space, material and energy footprint. No fun for me, because that's dealing with game engine engineering and not with the virtual world inside the game.

Just mentioning a different point of view, don't want to invalidate yours.

[spacedog]

I realized my fun is with designing factories that use the least ingame resources to produce their goods. Least amount of machines produce maximum amount of output with the least possible space. That's what I'm optimizing, and that's what I find big fun.

Yeah, this is where I'm at too. For me the fun is making an efficient factory (in the traditional sense). I like learning about the effect various design choices have on UPS, but maximizing UPS isn't the north star for my factory designs.

It's kind of a shame that UPS is what it is, to be honest. Trains are fun, and thematically they add a lot to a factory. Except apparently it's more UPS efficient to belt in resources from 40 km away over the course of 2 hours.

Same goes for nuclear power. You see all these monster 40K SPM bases, with a ludicrous array of solar panels hanging off to the side that's 5 times as big as the rest of the base. IIRC it was one of Stevetrov's bases where a significant chunk of the description about the build was dedicated to the logistics involved in constructing a solar array big enough to power the base. I understand the underlying UPS reasons, but clearly the game design intended for people to move on to nuclear power at those scales.

Anyway, off topic a bit. I appreciate all the points of view, and I get now that threads like these are about playing a specific facet of the game. Thanks everyone for taking the time to respond!

[Belter]

My second submission
https://factoriobox.1au.us/map/info/411 ... 4a91e874a4

I cannot download it, gives 404. Can you please re-upload it?

[Stevetrov]

My second submission
https://factoriobox.1au.us/map/info/411 ... 4a91e874a4

I cannot download it, gives 404. Can you please re-upload it?

It should be fixed now. same link.