Storage and lag: An experiment.

[SamuraiJones]

Hey peeps

Just ran an experiment and thought people might be interested in the results. So lately it seems like the popular storage method is dumping everything into a single pile via chutes/dispensors. I wondered if that was more cpu intensive than storage compactors, so I tried it. Based on what I'm seeing, it looks like storage compactors are better.

 

I didn't manage resources super carefully, so my hatches would've eaten some rocks and more regolith would've rained down (although my dupes couldn't access the surface so that wouldn't be a huge factor) while moving stuff from one location to the other, but dupe/critter count should be the same, as well as anything related to pathing.

 

As you can see, i've got almost 7000tons of minerals alone, so quite a bit of material.  In the big-pile approach, the pile was in a vacuum on a mesh tile, so theoretically it shouldn't be running any heat calculations on it. Also, the door was locked to maintain the air seal, so dupes didn't have access to the materials, which I also thought would improve performance.

 

Any thoughts? I'm kinda surprised by the results myself, and curious if anyone sees any issues with the approach.

 

(looks like the screenshot didn't include fps, you I guess you'll just have to take my word for it.)

Everything in storage compactors: 22fps (+/- 1)

https://steamcommunity.com/sharedfiles/filedetails/?id=1842611967

compact 22 fps.sav

Everything in a big pile (in a vacuum): 20 fps. (+/- 1)

https://steamcommunity.com/sharedfiles/filedetails/?id=1842611802

pile 20fps.sav

 

(Edit: Attached save files)

[SharraShimada]

This would make sense. 

Given the materials are all at a different temperature, they wont merge. So, not 100t big pile, but 5t 20°, 20t 50° and so on. But that means, the game has to calculate every pile, if there could be heat transfer. It can not, next pile. On a loop.

Stored in a compactor, the temperature gets even almost instant. From that point on, its just one 20k pile of stuff, calculated as one. 

In your case, ~350 calucaltions in compactors, vs maybe millions of small stacks. 

Did you made sorted compactors, or assorted? 

[Lifegrow]

I'd love to see a recording of the FPS change.

In my experience it makes zero difference.

 

[Lilalaunekuh]

2 hours ago, SharraShimada said:

In your case, ~350 calucaltions in compactors, vs maybe millions of small stacks. 

But at when you do a save/load cycle the game should merge the stacks.

=> After a save/load cycle should the vacuum pile contain up to 25t stacks.

 

@SamuraiJones Did you reload the map after dumping everything on a single tile ? [If not could you rerun your savegame and report.]

[SamuraiJones]

9 hours ago, SharraShimada said:

Did you made sorted compactors, or assorted? 

Sorted, if i understand you properly. A compactor had a single material, so one had only granite, another had only igneous, etc. At least for the bulk storage... there probably are a handful of old compactors around with multiple items, but not many.

 

7 hours ago, Lilalaunekuh said:

Did you reload the map after dumping everything on a single tile ? [If not could you rerun your savegame and report.]

I just ran it a second time to see if it made a difference... It did not appear to.

 

Further observations:

The pile approach seems to have a much more stable frame-rate. It's sitting at 21fps now (i maybe had more background stuff running before), swinging maybe between 20-22. The compactors seems to be averaging about 23, but it's swinging as low as 19 and as high as 28. Without formal datacollection, I'm not sure i'm confident that 23 is the average.

 

Btw, I'll add the save-files to the original post in case you guys want to try it on your pcs.