Performance optimization - post your saves

[Jap2.0]

Core count means nothing. It's limited by RAM latency which is limited by the speed of light and the distance between the CPU and RAM sticks on motherboards.

Patch notes 0.19:
- The speed of light was causing a lot of performance issues but we found a workaround to increase it 180%.

Only 180%? Man, the devs must be starting to get lazy.

Light is slowed down by gravity and you can make your CPU only have so much lighter with helium cooling.

Or you can use quantum entanglement on the cpu and the ram to have no delay in the communication, and then have infinite speedup and not worry about optimations ever again...

You know Rseding still would even if there was no need.

[Koub]

[Koub] This is fun and well, but please let's stop the lolesque offtopicness and refocus to save posting for performance optimization sake.

[c0bRa]

I've got a "robo only"-Base with about 175k robos (way too less for the things wo have to be done) in flight and many blueprints which should be done by the contructors atm... On my computer it runs with only 17 fps...
It doesn't hit the 60k on screen robos... Shall i post it anyway?

[DainPiediferro]

I've got a "robo only"-Base with about 175k robos (way too less for the things wo have to be done) in flight and many blueprints which should be done by the contructors atm... On my computer it runs with only 17 fps...
It doesn't hit the 60k on screen robos... Shall i post it anyway?

split the base in section, you'll have a massive increase of FPS
i doubled them from 30 to 60 splitting in section.

[PunkSkeleton]

Since 0.16 I noticed zooming out the map drops my UPS frpm 60 to 40 (was stable around 30 on 0.15). Prepare render takes 8ms when zoomed out and that's in mostly idle areas. Is this something normal because if I remember correctly it was way faster on 0.15.

[Rakshasa]

During 0.15 I had a game with an UPS of 20 on my rather beefy computer, very belt heavy in anticipation of 0.16. Lots of biters.

With the first experimental release I got back to 40 UPS, as expected. Added lots of artillery around the perimeter and fell back down to 25 UPS. Removed artillery, got back to 40.

Now, after a few updates and not that much expansion or changes it has fallen down to 25-30 UPS. I suspect it might have something to do with the biters as I've been using artillery wagons to clear out nests, and the biters just seem to hang around afterwards doing nothing.

Update: was down to 19 UPS, killed all enemy biters using command and now it is back to 40 UPS.

[Mimos]

During 0.15 I had a game with an UPS of 20 on my rather beefy computer, very belt heavy in anticipation of 0.16. Lots of biters.

With the first experimental release I got back to 40 UPS, as expected. Added lots of artillery around the perimeter and fell back down to 25 UPS. Removed artillery, got back to 40.

Now, after a few updates and not that much expansion or changes it has fallen down to 25-30 UPS. I suspect it might have something to do with the biters as I've been using artillery wagons to clear out nests, and the biters just seem to hang around afterwards doing nothing.

Update: was down to 19 UPS, killed all enemy biters using command and now it is back to 40 UPS.

Maybe the biters absorbed enormous amounts of pollution that now wander around for a long time? But I hope that pollution is not that expensive.

[pleegwat]

Artillery auto-targeting?

[Rseding91]

Artillery auto-targeting?

Artillery targeting has a fixed cost per artillery piece that's active (which isn't that high). Biter movement and attacking is what's expensive and will be triggered by artillery shooting at them.

[Rakshasa]

Artillery auto-targeting?

Artillery targeting has a fixed cost per artillery piece that's active (which isn't that high). Biter movement and attacking is what's expensive and will be triggered by artillery shooting at them.

When I added artillery to my border (before removing them again), they weren't actually shooting at anything.

So that seems not entirely correct.

http://rtorrent.net/private/railroad_104.zip

In the above save I have 14 UPS, and after killing all biters it jumps to 42 UPS.

[Lilly]

I've build a 1TW nuclear reactor, here's the save file. Even while it's being idle, the frame rate is already as low as 10fps. However, when turned on -- turn on the constant combinator that is on the hazard concrete, which you're standing when loading the file, such that the light above it turns green -- and after having spend 2 full nuclear fuel cells, the frame rate has dropped to only 3fps. Also the size of the save file will increase from 35Mb to 350Mb. I guess all the different fluid and heat levels are a lot harder to compress.

[ejg]

Artillery auto-targeting?

Artillery targeting has a fixed cost per artillery piece that's active (which isn't that high). Biter movement and attacking is what's expensive and will be triggered by artillery shooting at them.

When I added artillery to my border (before removing them again), they weren't actually shooting at anything.

So that seems not entirely correct.

http://rtorrent.net/private/railroad_104.zip

In the above save I have 14 UPS, and after killing all biters it jumps to 42 UPS.

I experienced the same and reported this as a bug, but was not deemed a bug. However, I think it was the frozen biter bug (biters paths were calculated but the biters never attacked because they were frozen). Should be better in 16.16.

[husky777]

Hello. ~~ I spent about 770 hrs on my last save(6900+ satellites launched), overall playtime is about 2000 hrs, active time – a bit more than 1000 hrs. I wanna ask developers to review my observations and ideas how to make the game optimization better. It’s the way for me to be grateful to developers’ team for the best game I’ve ever played.

1) I’m playing with maximum biters available, and I noticed that biters accumulate on unexplored, but already generated edges of map. This takes about 10-25 FPS/UPS. My suggestion is to make them more aggressive. This should prevent biters’ accumulation in numbers that may decrease FPS or UPS. The other option, as for me, is turning pathfinding off for the biters which either standing still or out of players’ sight (or on the unexplored part of the map). Generally, I call to revise biters’ behavior to the release, because current system is decreasing performance after 200+ hrs on save.

2) I’m playing on vanilla version most of my time played(i.e. without any mods and any console commands). Single “production efficiency” is pretty fast to research, so often I can’t put next one to the queue. This is leading to resources overflow. Also, turning next research on is leading to bots overwhelming the base, and – bigger base produce bigger lags. My suggestion is to add research queue, at least for infinite-grade ones. Infinite-grade researches can also toggle on automatically, i.e. “production efficiency 99” ended, and production efficiency 100” starts immediately. This simple addition must make production on large bases more stable and also this should decrease peak loading while releasing tons of bots and starting all the machines.

3) Artillery train was aiming and shooting in about 30 sec from 1st to 4th grades of “artillery range”. When 5th and 6th grade was researched, time to aim and shoot rose dramatically. Also FPS/UPS dropped a bit. Probably, aiming algorithm isn’t optimized enough

4) It’s impossible to build a facility to produce 2000 science packs per minute while FPS/UPS is higher than 30. Neither conveyers nor bots can’t resolve this problem. In my opinion, this is because game can’t effectively handle processor time and memory usage. The only thing working properly, as for me, is usage of video card.

5) Absence of ultimate-tier production machines, conveyers and energy generation in the late game (after rocket launch) is producing many lags, too. Large fields of machines and solar panels and kilometers of conveyers affects the performance seriously. As for me, this is (along with ineffective resources usage) the main reason of lags

6) Large facilities is also producing much pollution. More pollution requires more calculations, and this is leading to lags, too. I suppose it would be nice to have “green” alternative - expensive, but producing times less pollution than existing technologies.

7) The other thing is absence of quick and effective late-game items loading both into machines and storage. This is also leads to infrastructure overgrowth

8) Some people say it’s necessary to nerf bots. I think it’s not the acceptable for late game. Early- and mid-game bots’ nerf is OK, but I believe that late-game bots’ nerf will lead to performance drop. This may occur because of players who are used to bots and after nerf they will just build more bots. More bots = more lags.

9) “Compilation” of some previous points – the main problem of late-game is absence of other way to increase your production than building tons of production lines, solars, etc.

And also, please, check my list of suggestions which is aimed to improve late-game experience (after rocket launch and nuclear energy is available; this should also reduce lags):

a)Large chests or warehouses (as in warehouse mod)

b)Use electricity-powered loaders for quick items transfer to machines and storage

c) More effective and more expensive nuclear reactor (probably cold fusion or something like ITER, should be VERY effective and very expensive and complicated)

d) Introduce tiers for solar panels. Probably introduce wind electricity or hydro electricity. Existing solars are too expensive in terms of area

e) Introduce “green” technologies – air cleaning or machines wih reduced pollution

f) Introduce infinite resourses – to reduce map size. Probably rockets should “mine” resources

g) AAI vehicles mod is one of the best for me for both tweaking gameplay and reducing lags, for example, full clean-up of biters with small vehicle groups instead of using tons of turrets

h) Resource spawner overhaul mod is also a great tweak to map generation – rare but very dense (or even infinite) resource piles should reduce infrastructure difficulty in late-game which should lead to better performance.

[Rseding91]

1) I’m playing with maximum biters available, and I noticed that biters accumulate on unexplored, but already generated edges of map. This takes about 10-25 FPS/UPS. My suggestion is to make them more aggressive. This should prevent biters’ accumulation in numbers that may decrease FPS or UPS. The other option, as for me, is turning pathfinding off for the biters which either standing still or out of players’ sight (or on the unexplored part of the map). Generally, I call to revise biters’ behavior to the release, because current system is decreasing performance after 200+ hrs on save.

Pathfinding is not what makes biters expensive. Them moving is what's expensive - the game was built around non-moving entities being optimized heavily because that's what 99%+ of them do (trees, your entire factory, resources and so on).

3) Artillery train was aiming and shooting in about 30 sec from 1st to 4th grades of “artillery range”. When 5th and 6th grade was researched, time to aim and shoot rose dramatically. Also FPS/UPS dropped a bit. Probably, aiming algorithm isn’t optimized enough

Artillery targeting has a fixed cost per tick no matter the range upgrade applied.

4) It’s impossible to build a facility to produce 2000 science packs per minute while FPS/UPS is higher than 30. Neither conveyers nor bots can’t resolve this problem. In my opinion, this is because game can’t effectively handle processor time and memory usage. The only thing working properly, as for me, is usage of video card.

Yet people have still done it - with no mods - you simply need to build differently. "one big robot base" does not work here.

5) Absence of ultimate-tier production machines, conveyers and energy generation in the late game (after rocket launch) is producing many lags, too. Large fields of machines and solar panels and kilometers of conveyers affects the performance seriously. As for me, this is (along with ineffective resources usage) the main reason of lags

More stuff to process means more time taken to process it. The only exception being solar panels and accumulators - no matter how many of those you built they have the same O(1) cost.

6) Large facilities is also producing much pollution. More pollution requires more calculations, and this is leading to lags, too. I suppose it would be nice to have “green” alternative - expensive, but producing times less pollution than existing technologies.

Pollution only has a performance hit if it's over biters.

a)Large chests or warehouses (as in warehouse mod)

Large chests or warehouses are worse for performance than several small ones.

b)Use electricity-powered loaders for quick items transfer to machines and storage

Loaders are worse for performance than inserters.

c) More effective and more expensive nuclear reactor (probably cold fusion or something like ITER, should be VERY effective and very expensive and complicated)

d) Introduce tiers for solar panels. Probably introduce wind electricity or hydro electricity. Existing solars are too expensive in terms of area

Solar panels have a fixed O(1) cost no matter how many you build and area existing doesn't effect performance - only save size.

[Dooces]

Performace is amazing, relative to .15 anyway. Im sure there is still a lot more i can do to squeeze out some more UPS, but its great when you guys do it for me

heres my save for your consideration. I dont know how to use less insterters to achieve the same throughput. im considering circuit controlled inserters out of smelters on a timer, so they can always move 12 items per swing, instead of 1-2, but i dont know if that would be better or worse overall. having some option for inserters to only activate if they can move the full 12 at a time would be great!

Or change nothing. or change everything. imo the game has only gotten better with every decision made so far!

*edit, if smelters could make 12 items at a time, that cost 12x the mats, and take 12x longer... that would be nice ))

[mrvn]

Performace is amazing, relative to .15 anyway. Im sure there is still a lot more i can do to squeeze out some more UPS, but its great when you guys do it for me

heres my save for your consideration. I dont know how to use less insterters to achieve the same throughput. im considering circuit controlled inserters out of smelters on a timer, so they can always move 12 items per swing, instead of 1-2, but i dont know if that would be better or worse overall. having some option for inserters to only activate if they can move the full 12 at a time would be great!

Or change nothing. or change everything. imo the game has only gotten better with every decision made so far!

*edit, if smelters could make 12 items at a time, that cost 12x the mats, and take 12x longer... that would be nice ))

I'm thinking about making a mod that has every furnace and assembler building as a double sized entity (so a 2x2 furnace becomes 4x4) that costs 8 times as much (twice as high too so 8 times the volume) and handles all recipes in batches at 8 times the speed. A batch size of 12 would mean each recipe takes 1.5 times as long to produce 12 times the output at 8 times the cost to build. You save a little space but otherwise the balance shouldn't be thrown off by this. But you would get your stack inserters to always pick up 12 items.

[foodfactorio]

Hi can i check if the game has changed anything regarding the number of simultaneous sounds from v0.15 to v0.16 (experimental)?
it might not be due to that, but usually i get good fps/ups (60/60 unless zoomed out about 30/60) and a lot more often now then before, when i walk near multiple buildings, it freezes for about 10 seconds, and then the game recovers audio again and i can continue playing, but its happening more often now)

[RainbowAndy]

Last few games I'm experiencing severe FPS drops to under 20. The factories aren't very large, definitely not mega-factories, but I do use BobsMods and SpaceX.

Computer is;
Intel Core i7-6800K CPU @ 3.40GHz
32GB installed RAM
64 bit processor and operating system
NVIDIA GeForce GTX 1060 6GB

Thanks in advance as Factorio has been my favourite game since first bought!

[Rseding91]

Last few games I'm experiencing severe FPS drops to under 20. The factories aren't very large, definitely not mega-factories, but I do use BobsMods and SpaceX.

Computer is;
Intel Core i7-6800K CPU @ 3.40GHz
32GB installed RAM
64 bit processor and operating system
NVIDIA GeForce GTX 1060 6GB

Thanks in advance as Factorio has been my favourite game since first bought!

It's because you have a bazilion biters on your map and they're all active near your borders waiting for the game to let them attack. You've simply gone beyond what the game can handle.

[foodfactorio]

hi, there is a mod here that might be able to help?
https://mods.factorio.com/mod/BitersBegone
i have not needed to use it myself but if you make a backup first, maybe that can help free up some fps, and then you can disable it again to still get a few biters spawning?

(btw i tried launching the game with my onboard intel gfx card, and the hanging/almost crashing issues went away. (those were with the nvidia card, which does have a lot more fps/ups than the intel, but as long as i dont zoom out too much, i can still get at least 30fps or more with the intel)