Page 1 of 1

[2.0.15] Discrepancy in Updage time usage breakdowns

Posted: Thu Nov 14, 2024 11:58 pm
by fwyrl
I am once again looking at my Update Time because although the 2.0 update helped a lot, adding more factories has eaten many of the structural gains. I can optimize my factory to improve the UPS further, but that's where the time breakdown comes in.

As I'm looking through it, I've found that the general time usage's listing of "Entity update" is always 10-20 ms greater than the "Entities total" listing in the entity time usage. What is it that's not being reported here? Or am I misunderstanding this?

Snapshot of Entities Total and Entity Update
Snapshot of Entities Total and Entity Update
11-14-2024, 18-48-14.png (314.72 KiB) Viewed 310 times
Snapshot of Entities Total and Entity Update
Snapshot of Entities Total and Entity Update
11-14-2024, 18-48-26.png (230.73 KiB) Viewed 310 times
Snapshot of Entities Total and Entity Update
Snapshot of Entities Total and Entity Update
11-14-2024, 18-51-54.png (234.75 KiB) Viewed 310 times
Snapshot of full time breakdown
Snapshot of full time breakdown
11-14-2024, 18-49-09.png (1.83 MiB) Viewed 310 times

Right now I'm looking at my biggest UPS savings being eliminating biters and therefore being able to eliminate my turrets, followed by looking into why my power network is using so much performance. Then overhauling my factory to use the new buildings, the new fluid system, fewer inserters, no chests, etc, but there's some 10+ ms drain I can't find, which is larger than the largest drain I know of (turrets and inserters at 7ms both)

Re: [2.0.15] Discrepancy in Updage time usage breakdowns

Posted: Mon Nov 25, 2024 8:01 pm
by EnerJi
Hi, I'm curious whether you figured out what was going on and / or how to interpret the missing time. I'm interested in learning how to better interpret the debug statistics for performance troubleshooting purposes.

Re: [2.0.15] Discrepancy in Updage time usage breakdowns

Posted: Mon Nov 25, 2024 10:53 pm
by Rseding91
All of logistic and construction robots will be included in the "overall" time usage but not the per-entity breakdown. Additionally, turrets, beacons, and some other smaller entities will also work that way.

It's almost always obvious looking at any save file what's taking the time in the save. The most common thing is "the thing you built the most of". That will be one of:

* Robots
* Inserters
* Belts
* Trains

Re: [2.0.15] Discrepancy in Updage time usage breakdowns

Posted: Mon Nov 25, 2024 10:57 pm
by BraveCaperCat
Rseding91 wrote: Mon Nov 25, 2024 10:53 pm All of logistic and construction robots will be included in the "overall" time usage but not the per-entity breakdown. Additionally, turrets, beacons, and some other smaller entities will also work that way.

It's almost always obvious looking at any save file what's taking the time in the save. The most common thing is "the thing you built the most of". That will be one of:

* Robots
* Inserters
* Belts
* Trains
Usually, and in my base, I don't have any trains - but I've been to (but not automated)
Vulcanus
,
Gleba
and
Fulgora
.

Re: [2.0.15] Discrepancy in Updage time usage breakdowns

Posted: Tue Dec 03, 2024 5:31 pm
by fwyrl
Rseding91 wrote: Mon Nov 25, 2024 10:53 pm All of logistic and construction robots will be included in the "overall" time usage but not the per-entity breakdown. Additionally, turrets, beacons, and some other smaller entities will also work that way.

It's almost always obvious looking at any save file what's taking the time in the save. The most common thing is "the thing you built the most of". That will be one of:

* Robots
* Inserters
* Belts
* Trains
I see a "Turret" category in the per-entity breakdown, is that not the whole story for turrets?

Will look into these on my main base (delete them all in editor and see what happens), and go from there. Thank you for your time :)

Are biters, stompers, etc under the "Unit" section in entity time breakdown? I suspect the number I have would exceed 2 ms/t, but maybe they're a lot more efficient than I thought.

Edit: after doing some testing, the discrepancy was a combination of several things, but Turrets were actually costing about twice what was in the entity breakdown - 20-some ms/t instead of about 9-10. Robots, Roboports, Radars, and beacons made up the rest, in about even amounts; (<5 ms/t each)
Surprisingly, all those biters were using under 5 ms/t!