Road traffic is a known fps killer, but how much impact has it exactly? I tested it out with my setup - a vanilla install of Prepar3D v4.1. with medium high graphics settings. I used the avatar and placed him somewhere in downtown Tokyo. Weather and other AI traffic was turned off to isolate the fps impact from the cars as much as possible.

Of course, Tokyo has extreme levels of road traffic, but nonetheless you get an idea of just how much of an impact this setting alone can have on the sim. The biggest surprise for me was how much impact just 5% road traffic had. At 50% road traffic this scenario lost about 50% of its fps, and the sweet spot for me seems to be 10%. 

Many thanks for your analyses

 

One question: do you have enabled shadows? (in the meaning that road traffic generates shadows?)

Yep, I have simobjects shadows enabled (as can be seen in the screenshot ;)).

1 hour ago, simmerhead said:

as can be seen in the screenshot

 

was not 100% sure when looking at the screenshot...

thanks for answering

Thanks for this. Its why I have my road traffic at no more than 10%. I just want to see some cars going past regularly and that seems to be enough. Packed roads does not add to the immersion for me.

So going from 5% to 10% impacted FPS by 2 only? Interesting. I guess I should go back to 10 from 5 then. Wonder how ships and leisure boats would affect FPS in the same scenario. 

I think having shadows on the cars might have caused a lot of the FPS hit, how do the figures look at higher densities with AI car shadows disabled?

Hi again. I loaded the test scenario and did a few comparisons (I did a slight change in my settings, so the fps baseline is a tad bit different from the table above). 

Anyway, here are the numbers (I load up the scenario and let things settle for a 2-3 minutes before reading the average fps):

50% road traffic, with Simobjects cast shadows enabled:          28,1 fps
50% road traffic, with Simobjects cast shadows off:                   28,3 fps (No major hit. I run shadow settings @ medium + medium)
50% road traffic + 100% ship traffic:                                           27,9 fps
50% road traffic + 100% ship traffic + 100% small boat traffic:  27,4 fps

Remember, these settings are relative to other settings and the location you're flying. But for airliner flying, which generally happens between larger cities, I find the numbers useful. Ship traffic probably has more impact in other places, especially around Seattle where there is a lot of ferries as well. The small boats are spawned on the fly as you fly over them, and I assum their impact is quite limited overall as well, although it will add additional workload to the CPU. 

Great info! Thanks very much for the analysis...

Thanks for testing that out. I guess it’s best to keep all 3 at 10% be done with it. 

For ships and boats 10% seems to be more realistic anyway or else it spawns a whole bunch of them in one small area instead of spread out in the ocean lol. 

Thank you for taking all that time to test and report. I will keep mine at 0% 

I prefer "Ships and Ferries" to be @ 100% - I like having the larger boats around - unless I'm using Henrik Nielsen's excellent Global Ship traffic package.

I keep "Small boats" @ 10-20% during summer, and forget about them during winter. 

 

Ships and ferries don't spawn like the small boats do. They act very much like aircraft traffic with ships following routes coded into a bgl file stored in the Scenery/World/Scenery folder. 

Thanks for your testing .

2 hours ago, simmerhead said:

I prefer "Ships and Ferries" to be @ 100% - I like having the larger boats around - unless I'm using Henrik Nielsen's excellent Global Ship traffic package.

I keep "Small boats" @ 10-20% during summer, and forget about them during winter. 

 

Ships and ferries don't spawn like the small boats do. They act very much like aircraft traffic with ships following routes coded into a bgl file stored in the Scenery/World/Scenery folder. 

Doesn't that mean they use more overhead like AI Traffic? I don't mind keeping them high if they don't eat into performance.

1 minute ago, TheBoom said:

Doesn't that mean they use more overhead like AI Traffic? I don't mind keep them high if they don't eat into performance.

The default ship traffic is quite sparse, so I don't think you will notice much performance degradation.