Post your AI benchmark score

MrMorph

Chieftain
Joined
Oct 8, 2015
Messages
10
Hi there,

I've recently discovered AI benchmark, that is measuring how much time ai turns take. It would be great, if we could post here our results, as it might be useful for those that are thinking about upgrading their computers.

You can find Benchmark in game's main menu.

Cpu: i5-2700k
GPU: AMD Radeon 6800
Ram: 8gb
Result: 27 seconds
 
Last edited:
Good idea, but it will vary a lot with the number of civs and CS,
whether you play with barbarians on or off, etc.
Do you want results for 33 civs and 28 city-states on ludicrous (200x100)
maps, or just standard maps?
 
Good idea, but it will vary a lot with the number of civs and CS
I guess he is talking about the Benchmark test from the main menu.

CPU: i7-7700k
GPU: GTX 1070
RAM: 16GB DDR4 3200Mhz
Result: 13.03s

With the same system setup, I've got 16.85s, a few months ago before R&F. So almost 4 seconds it's a good improvement.
 
Last edited:
CPU: i5-7600k
GPU: GTX 1070
RAM: 16 GB
Result: 15.99 seconds

Edit: Seems to vary from 13 to 16 seconds. Might be because I run in windowed mode.
 
Last edited:
Thanks for suggestions guys! I have edited main post, so it might be more clear for the others.
 
CPU: i7-6700k
GPU: GTX 1080 x 2
RAM: 16GB DDR4 2133mhz
Result: DX11: 12.78 seconds
DX12: 12.99 seconds
 
CPU: i5-3350P @ 3.10 GHz
GPU: AMD Radeon R9 290 (4GB)
RAM: 8GB
Result: 20.77 (second run)

Turn 1-5
20.19209
21.24996
20.96777
18.18986
23.14893

Turn 1-5 (second run)
20.31162
21.26662
21.00773
18.15358
23.099

It's good to know that when I finally get around to upgrade my system I will get to take off a few seconds from the turn times.
 
CPU: i7-3770k
GPU: GTX 970
RAM: 16 GB RAM
Result: 15.49

I'm surprised it's this good because my CPU is quite old. Oh well, I knew there was no reason to upgrade.
 
CPU: i5-6600k
GPU: GTX 980 ti
RAM: 16 GB RAM
Result: 19.48

A bit disapointed...I'd like to be under 15s. Any tips to improve ?
 
maybe i'll try this when I get home. Didn't even know it was a thing. It seems faster post-expansion to me but that's just based on memory and no numbers.
 
CPU i7-4790K 4Ghz
GPU GTX 970 4GB
RAM 24 GB
Result 14.39
2nd Run - Graphics set to max performance
Result 14.08
3rd Run - Graphics set to max quality
Result 14.53
 
Last edited:
CPU: i5-6600k
GPU: GTX 980 ti
RAM: 16 GB RAM
Result: 19.48

A bit disapointed...I'd like to be under 15s. Any tips to improve ?
Looks like i7 hyperthreading is used well in CIV 6.

Doesn't the Benchmark tool use your current graphics settings to perform it's tests? So two people with the same CPU & GPU, but different graphics settings would have different turn times.
That's a good question actually. For the sake of a comparable benchmark I'd assume it's the same settings for everybody but that might be wrong.
 
For the sake of a comparable benchmark I'd assume it's the same settings for everybody but that might be wrong.
It's been about a year since I remember reading about the graphics settings, but the more I think about it I believe I'm mistaken. It was the Graphics Benchmark that depended on the graphics settings not the AI Benchmark.

CPU: i5-4690k (4.5GHz o/c)
GPU: GTX 1060 SSC 6GB
RAM: 16GB DDR3 2133MHz
API: DX12
Result: 13.57
-----------------------------------------
Graphics Settings: High
Average fps: 58.76
 
i7 4790k (at 4.0)
Radeon rx480 4GB
16GB ddr3 2133
14.07

I thought I overbuilt when I put this together a few years ago but I guess it's still holding up (aver an overdue video card change).
 
Last edited:
i7 3820 @4.8
16GB DDR3 1600
1080Ti @2.05
14.28s

EDIT: As it seems that graphical settings impact AI benchmark, here are my settings : All maxed, 2160p
 
Last edited:
CPU i7-4790K 4Ghz
GPU GTX 970 4GB
RAM 24 GB
Result 14.39
2nd Run - Graphics set to max performance
Result 14.08
3rd Run - Graphics set to max quality
Result 14.53

To answer the question about graphics settings effecting the results I ran two more tests. The first (2nd run) I went with highest performance possible (all settings turned down or unchecked). The second (3rd run) I went with max quality (all settings checked and highest option selected). My original test was with animated leaders [medium], terrain quality [medium - performance optimized], and high-quality visual effects [low]. All sets had vertical sync checked.

As you can see between the max performance and max quality I only had a difference of 0.45 or an increase of 3.09% when running on the highest quality possible.
 
Top Bottom