#[SPZ2-4758] [0.1.1] - Overflow Splitter: Inconsistent Behavior

44 messages · Page 1 of 1 (latest)

drifting vigil
#

Bug Report

Similar to https://discord.com/channels/1000343719314198548/1327935037181792297, the Overflow Splitter is triggered randomly, even though there is no visible Overflow to be seen.
-# As seen on the attached GIF.

I've only noticed this issue when placing the Overflow Splitter right after the Belt Catcher, but not after any Conveyor Belt (as of writing this).
-# I'm currently not aware of any other combinations that could potentially trigger this exact issue.

The Throughput / Belt Speed should be capped at 180/m, yet it appears to randomly oscillate between 181/m and 182/m, and it also appears that the Overflow Splitter itself is working at ~101% capacity.
-# As seen on the attached Screenshot. (Perhaps just a visual error?)

Sidenotes

  • I acquired the Shapez 2 Supporter Edition.
  • I've encountered this issue prior to the release of the v0.1.1 Dimension Update, sometime during May.

Settings

Computer Specifications

(as seen in the Game's Settings on the top-right corner)
-# I'm able to run the game at roughly around 30-60 FPS for the most part.
AMD Radeon RX 6600 XT
Direct3D11
ATI (Direct3D 11.0 [level 11.1])
8,120 MB VRAM

MS-7C91 (Micro-Star International Co., Ltd.)
AMD Ryzen 9 5900X 12-Core Processor
24 x 3700 MHz
32 MB RAM

General

Simulation Threads: Auto
Simulation Details: High
☑ Parallel Rendering

Graphics

Borderless Window Mode
3840 x 2160 Resolution
VSync: Off
60 FPS Limit
☑ Reduce Input Latency

Antialiasing: Off
Ambient Occlusion: Extreme
Shadows: Extreme
Shaders & Effects: Extreme
Space Background: High
Space Platforms: High
Buildings: High
☑ Chromatic Aberration
☑ Vignette
☑ Anisotropic Filtering
☑ Post Processing
☑ Interface Effects

💠 Feel free to ask any questions, and I'll do my best to respond.

shapez_2_2OZduAxSZv.gif shapez_2_j3i82frf0E.png
left vapor
#

I've seen this problem incorrectly divert 20-30% of shapes.

drifting vigil
#

I forgot to mention that this happens to every 2nd to ~10th shape.

elfin abyss
#

Thank you for your report @drifting vigil ! 😄

I seem to have some trouble reproducing the issue on my side. Would you be willing to share your save affected by this issue?

left vapor
#

Its a little hard to reproduce. Usually packing multiple belts of shapes onto a space belt lane that then goes into a catcher then overflow splitter makes it happen pretty reliably.

elfin abyss
#

Thank you for the response @left vapor ! It seems that what causes the issue is directly placing the Overflow splitter after a catcher, and does not occur when there's at least 1 belt between them. I'll write a ticket right away into our internal system for further investigation 🙂

left vapor
#

I have absolutely seen this happen with multiple belts between the catcher and splitter though... I couldn't reproduce that.

elfin abyss
#

There has been some refactoring on the belt speed that might have fixed that occurring, but that might have missed this edge case

drifting vigil
#

Upon further messing around, it continues to be even more inconsistent, as at times it will behave as you expect, and at times it won't.
Here's a poorly-made recording of me cutting & placing the platform:

left vapor
left vapor
drifting vigil
#

This occurs regardless of whether there's a backlog or not. Although I'm not sure what else I could do to better replicate the issue.

left vapor
#

yes it happens without a backlog. (but a little less frequently in my experience)

drifting vigil
#

Even reloading the save file seems to have a 50% chance of it either working as expected, or exhibiting the issue.

elfin abyss
#

I believe the length of the main path and/or split path might have to do with the frequency of the pathing, but that's pure conjecture. In any case, I'll write a ticket for the splitter having much more speed when placed directly next to catchers (which causes the speeding up) and hope the devs can catch any other redundancies 🙂

Thank you for all the info you found, I'll add it to the ticket for the team to have all available in one place! 🙏

left vapor
#

I suspect this is simply a problem with space belt to catcher connections and the overflow splitter doesn't respond to it gracefully.

elfin abyss
#

[SPZ2-4758] [0.1.1] - Overflow Splitter: Inconsistent Behavior

left vapor
#

I'm getting overflow splitters trashing shapes for no reason outside of this niche scenario in 0.1.1 (No space belts involved at all). This bug should be pretty high priority. I can't reproduce it yet though.

left vapor
modern marsh
#

i have this
video below

left vapor
#

This looks like expected behavior. The painter delays 1/4 of the shapes so that they interrupt the 3/4 causing them to pause (extremely) briefly (which makes the belt stopped)

modern marsh
#

i did not expect it 🙂

left vapor
#

yeah its really annoying that it works that way

modern marsh
left vapor
left vapor
modern marsh
#

But maybe both can be solved with the same fix, i think it would solve these if the shapes travelled faster in intersections.

#

(treating the overflow splitter as an intersection too)

#

(but someone might use fake intersections to achieve lower latency)

opal rune
modern marsh
#

an intersection can store more shapes when there is a blockage, so maybe it is only a little bug

image.png
left vapor
#

its now every other shape (in the scenario in my BP)

left vapor
scarlet zodiac
#

Just do double check, in alpha 3 it still works correctly if you place a belt before, right?

unborn silo
unborn silo
#

[SPZ2-5688][SPZ2-4758] [0.1.1] - Overflow Splitter: Inconsistent Behavior

#

[SPZ2-4758] [0.1.1] - Overflow Splitter: Inconsistent Behavior

#

The cause isn't the overflow splitter. The belt catchers/launchers can take shapes faster than 100% if space belts are supplying faster than the belt catchers/launchers can receive.

proven quartz
#

This issue is still occurring in 1.0.3-RC3

proven quartz
#

It was diverting 15-35% in my massive Manufacturing world

I opened up a blank sandbox to see if it would go away without CPU lag, but it still diverts about 11%