Performance survey.

[m04sWYEGzS]

I can't imagine any way we could "skip ahead". Even if your region isn't actively doing a lot of stuff, all the other countries are doing stuff, maybe even wars. Our goal is for players to have something worth doing "every game day". The only way we get to 1sec/day is to reduce game complexity, which is at the heart of what makes the game so engaging.

I'm frequently in a situation where i'd like to skip a couple of days. units lost in the Serbian Tundra, waiting for resupplies, no ongoing wars and only researching/supplying, ruler of the world with an unfinished research tree, etc. You definitly shouldn't reduce the game mechanics for this, but you shouldn't aim for a "10s on fastest is sufficient". Why would you even implement slower game speeds if the fastest is the only "playable" one? it should be more like:
Fastest: 1s
Very Fast: 10s
Fast: 30s
Normal: 60s
Slow: 90s
Slowest: 120s
If the slowest speed would take me more then 2min, why shouldn't I simply pause the game if it really is necessary to do so much that 2min is not enough?

[rlame]

Current game build.
Ryzen 5900X, RAM 32Gb 3400mhz
first day 14.xx seconds (did not catch precise)
10 days 1min 40.68s
10 days average = 10.068s

[Balthagor]

anyone tried this speed test lately?

[ImperiorIL]

I loaded it, didn't move the camera, or change the zoom. Max settings at 4k. Speed 5, gave it a few seconds to load before unpausing.
I tried it twice, got somewhat different times with noticeable speedups seemingly at random.

The first time, I ran up to the 7th, 11 days, and got 2:52 or ~15.6s per day.
The second time I ran up to the 6th, 10 days, and got 2:50 or 17s per day.

My computer is a I7-5820k overclocked to 4.3GHz with 6C/12T. 32GB of Ram and a GTX 980ti.
It is worth noting my storage raid is really slow though. Else, running on Steam with the Overlay.

PS: on a save I started in 1949 as New Zealand, I noticed it takes about 40-60 minutes per year by the mid 1980s.

[ImperiorIL]

Why would you even implement slower game speeds if the fastest is the only "playable" one?

Whenever I'm microing units post 1970s, so jets and modern tanks/mechanized inf, I find myself often running in Slow-Fast. Even Very Fast is too fast to micro units.
In Fastest, it is pretty much impossible to micro units. But yes, it is the speed that I stick to 99% of the time, I only change it when I'm personally controlling units.
Very Fast is currently a great speed for broad group unit moviments. Normal-Fast is good for controlling individual stacks of units.

So, no, I don't think the only playable speed is fastest. That's like saying the only playable speed in Europa Universalis, Victoria, CK, and HOI is speed 5.
It really depends if you are half heartedly managing your economy and letting your minister control your army or if you are actually microing stuff.
If the game is not [yet] lagging, speed fastest is too fast for diplo even, you cannot propose stuff before prices change, commodities change, etc.

I would argue "very slow" is completely useless, currently. And that I wouldn't really miss it if slow was gone too.
Sincerely, if the game ran consistently a day in 10s at worst, and faster than that most times during peace, I'd already be quite happy.
The absolutely awful AI, unit pathing, unit slow reaction, supply simulation, and gigantic 16km hex size, are much more painful to the game than the run speed.
If any optimization was made that made it more multithreaded or run faster with AVX-256 or something, using the extra performance to solve the above would improve the game a lot more than just having it run faster.

[Balthagor]

Thanks for the feedback!

[Balthagor]

New computer (cause my video card died)

Ryzen 3400G, 16GB ram, onboard video (cause video card prices are stupid)
First day, 30sec, 10 day average 22.5sec

[evildari]

i7-950@3.07GHz/16GB/GTX 1080 - 29sec first day, 4:25 to start of Sept 6th (26.5sec average)
3440x1440 Graphics Resolution HIGH, everything else rather off, including music / sounds

previous test was on my old rig from early 2011 with WIN7.
zoomed in the southern atlantic with no units in sight.

Now did the same on my new machine running (customized light weight) WIN10:
i7-9700K@3.60GHz/32GB/RTX2070Super - 16sec first day, 2:12 to start of Sept 6th (13.2sec average)

so a roughly 10times higher model number after 9 years, yields just a double speed increase.

Update: also run the game on same new system with WIN7 (yes i hate WIN10 [_]B that much):
15sec first day, 1:59 to start of Sept 6th (11.9sec average) , ran a second time just to get sure - even 1:58 seconds...
another 10% saved

Same test - same machine - just other OS: WIN11 updated to current date - except some intel sm-bus or chipset drivers dated 1968
17.7sec first day, 2:35 to start of Sept 6th (15.5sec average)

also at mid-zoom level over europe fps gets sometimes one-digit - but usually is around 12-14 fps.

[Balthagor]

bump

[C0dingschmuser]

2560x1440, everything on max, win10
R9 5950x
RTX 3090
32GB ddr4@3733CL15

13s first day, 1:35min for 10 days (9.5s avg)

[DesTryke]

i5 8600k 4.4 GHz
DDR4 32 GB 3000 MHz

19s first day, 2:30min for 10 days

[arakan94]

Updated my system:
Ryzen 7 5800X
32 GB DDR4 3000
Radeon 6800XT

1st day: 16.10s
10 days: 2:05 = avg of 12.5s/day

FPS is still atrocious though - just 9-10 FPS on max zoom out :/ (weather, ownership, units enabled)

sru fps.png

If I disable all overlays (no weather, no ownership, no units), I get 23 FPS on max zoom out. I do really hope that performance will be much, MUCH better in the SR2030.. Heck, all I want is SRU with some UI improvements and good performance..

[arakan94]

Updated my system again - switched to Ryzen 7 5800X3D

1st day: 13.4s
10 days: 1:43 = avg of 10.3s/day

Big cache is really helping with those calculations (the CPU runs at cca 300 MHz less than 5800X). Would be really interesting to see performance with 7800X3D...