Factorio Forums

www.factorio.com
https://forums.factorio.com/

UPS optimized 12-beacon smelting - w/minimal circuits

https://forums.factorio.com/viewtopic.php?f=193&t=100763

Page 8 of 10

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Fri Mar 04, 2022 8:13 am
by disentius
There is no difference (anymore) between normal and filter inserters ups-wise.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Fri Mar 04, 2022 9:26 pm
by Belter
That was my point as well - but I'll test it just to be sure.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sat Mar 05, 2022 11:16 am
by disentius
Good point about the miners being in range of the beacons. I will adjust the saves and run it again.. thanks!

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 9:38 am
by Belter
Belter v6.99
- Tried to optimize the layout using the 32 / 128 timing
- My tests showed a little gain on filter inserters, so I've switched to them

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 10:00 am
by Belter
Benchmark - Non-DI BPs
- no mods
- 500 lanes, 1 circuit controller
- 4 runs x 25.000 ticks per save
- drills: mining-productivity level: 2000, no modules, not in beacon range
Saves
- Belter v6.99: my latest attempt, minimal # of belts
- DaveMcW v6: Dave's last BP with a different timing
- Belter v6.98: exposed as possible output belt
- Belter v6.97: exposed ore belts
- Belter v6.96: exposed ore belts, stack inserters, same layout as 6.97
- common for all: 144 cycle time, timing w/2 signals, separated, 2 x half lane ore feed, filter inserters unless stated otherwise
- all saves attached - Saves.Zip
Results, conclusion

- See charts and results.csv attached
- v6.99 beats DaveMcW's v6, but the difference is close to 1% - not significant.
- Dave's layout uses 1 less beacon
- transport lines times are better for v6.99, but entity updates are better for Dave's
- filter inserters ARE a little bit better, difference is under 1% but still consistent (6.96 vs 6.97 - that's the only diff)

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 10:16 am
by Belter
I hope that the slightly different drill layout on the left and the infinity chest positions on the right does not have a significant effect on the results.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 11:50 am
by disentius
We are in "our tools are not good enough for a meaningful distinction" country now.
I wanted to know the consistency of the results, so I took the same save, and copied by saving them under a different name, and ran the benchmark.
(the only difference between the saves is that the time of save (gamestate) is different.
all saves have a slightly different size -> the moment of saving matters.

test done on an "idle" (AFK) computer.

Results:
2022-03-06 12_41_58-Window.png
2022-03-06 12_41_58-Window.png (101.33 KiB) Viewed 6106 times

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 12:06 pm
by Belter
disentius wrote: ↑
Sun Mar 06, 2022 11:50 am
 We are in "our tools are not good enough for a meaningful distinction" country now.
I think 1% is a good number for accuracy. I'll run these just for science. Can you ran the top 2-3 above? You have 2x UPS than on my potato computer.

Update: I've run all 5 times, they're exactly in 1% range.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 12:41 pm
by disentius
To late:) Already ran another test.: clones from the same save.

Results:
2022-03-06 13_31_42-Window.png
2022-03-06 13_31_42-Window.png (42.38 KiB) Viewed 6076 times
My PC:
2022-03-06 13_33_45-Window.png
2022-03-06 13_33_45-Window.png (21.87 KiB) Viewed 6076 times
Files attached.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 1:42 pm
by disentius
Ran your top 3 saves.
Made 1 modification. i saved them on paused, with the timers on tick 1.
results:
(I calculate difference % on column execution time, since that is the exact value. Effective UPS is a calculated field)
2022-03-06 14_35_20-Window.png
2022-03-06 14_35_20-Window.png (49.74 KiB) Viewed 6051 times

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 2:28 pm
by disentius
A very interesting UPS comment by boskid:
viewtopic.php?p=562810#p562810

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 4:27 pm
by Belter
TY for running the saves!

Clearly the results are not consistent.
Update: I've scaled Dave's UPS to 100% and show both benchamrks at the same graph:
DisBeltBenchmark2.png
DisBeltBenchmark2.png (50.19 KiB) Viewed 5968 times
prev attempt, same numbers
disentius wrote: ↑
Sun Mar 06, 2022 1:42 pm
 (I calculate difference % on column execution time, since that is the exact value. Effective UPS is a calculated field)
Update: I think I misunderstood your point, let's not go there.. Eff UPS is calculated from Exec time, agreed. Also there is a rounding in the script. I prefer using UPS as that is what we optimize for here

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 7:57 pm
by Belter
I was able to gain +7% but.. cheated :D - used direct insertion and ore fields in the BP.
New thread for DI?
I'm considering opening a new thread:
- do what you want in the editor e.g. place ore
- produce 512 lines of plates to express loaders + infinity chests places outside your BP (so 8 belt constructs can be cloned easier)
- +same guidelines as we concluded here (vertical setup, chunk aligned, one controller, no mod benchmark, 5 x 25k runs)

For me this thread should not be about DI from minig drills insdie the BP.

Not sure who would be interested to continue w/DI. Let me know. Then I'll move the BP below there!

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 06, 2022 8:04 pm
by Belter
Belter v8.1
- DI of ore to smelters, needs ore patches inside the BP
- 10 beacon setup, cycle time is 168.42 ticks -> using 176 as 32 * 5.5
- further idea: with minimal effort, separate signalling / separate wiring is possible
- further idea: fully exposed belt is also easy to do
- further idea: signal optimizaiton might be also possible
- problem: not easy to clone this BP due to the overlapping minig drills
- problem: I see that the poles are inside the drills, will fix that in the next version (clone issue, my bad, poles to go centered vertically next to the beacons)
Belter v8.1.png
Belter v8.1.png (1.51 MiB) Viewed 5949 times

First results - clone issue to fix

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Mon Mar 07, 2022 2:46 am
by flame_Sla
Belter wrote: ↑
Sun Mar 06, 2022 8:04 pm
Belter v8.1
- DI of ore to smelters, needs ore patches inside the BP
the inserter picks up ore from the ground?
place the chest

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Thu Mar 10, 2022 10:27 am
by flomz
One belt in versions 6.98 and 6.99 seems to have ore on the wrong side.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sat Mar 12, 2022 3:29 pm
by Belter
flame_Sla wrote: ↑
Mon Mar 07, 2022 2:46 am
 the inserter picks up ore from the ground?
place the chest
Noted, TY! Will fix it in the next BP.
flomz wrote: ↑
Thu Mar 10, 2022 10:27 am
 One belt in versions 6.98 and 6.99 seems to have ore on the wrong side.
I beleive the UPS gain is to have 1 lane only, side not relevant. Will test this sometime.

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sat Mar 12, 2022 3:51 pm
by Belter
Updated: resolved, thanks DaveMcW! Forgot that I've added steel chests of the fast as suggested above by flame_Sla...

Testing different variants of my DI BP to get the highest UPS. See the 2 attached saves - they look identical. But the UPS difference is significant :shock:
problem was

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sat Mar 12, 2022 11:39 pm
by DaveMcW
Your counting script runs into console flood control, so it is not accurate. Try this version:

Code: Select all

/c for _, building in pairs({"arithmetic-combinator","beacon","constant-combinator","decider-combinator","electric-energy-interface","electric-furnace","electric-mining-drill","express-loader","express-transport-belt","express-underground-belt","medium-electric-pole","productivity-module-3","speed-module-3","stack-filter-inserter","stack-inserter","steel-chest"}) do 
  game.print(building .. " " .. game.player.surface.count_entities_filtered{name=building}) 
end
Spoiler

Re: UPS optimized 12-beacon smelting - w/minimal circuits

Posted: Sun Mar 13, 2022 9:24 am
by Belter
Belter v8.16
- direct insertion BP optimized
- uses perfect timing w/6 signals on 1 wire
- I've tried to put speed modules to drills, a bit worse UPS. interesting. So left them empty
- UPS gain: the 2 middle drills put ore to a chest vs ground
- exposed (not underground) belt possible, but not better UPS-wise

All times are UTC
Page 8 of 10

Powered by phpBB® Forum Software © phpBB Limited