All benchmarks are done with the 126.96.36.19977 Open Beta from Oct. 16th 2018.
The track used is very basic, I started a freeflight in Caucasus in a SU-25T and just flew around. Mostly low altitude in the mountains. So in most scenarios, CPU will be used more. i.e more units, more flying over cities etc.
Settings for both SteamVR and DCS. please note that the combined resolution for both settings (SteamVR scale and DCS PixelDensity) are roughly comparable to DCS PD 1.45.
Stock DCS tests
Let’s start with the results of the stock DCS.
in stock DCS it seems at first glance that the CPU is not the bottleneck, the numbers of the stock CPU with 2133 memory are slightly lower but nothing meaningful.
Now it’s time to plot the data.
Looking at the plots shows very little difference, tests are pretty consistent without many surprises sure, stock numbers are slightly lower, but not by much.
Modded DCS tests
With the mod in play, picture changes a little bit.
With the shaders mod, the load on the GPU is lowered, causing it CPU to be more of a bottleneck and here we can see that the memory speed and “Sync all cores” settings are bringing up the FPS count.
going up from there to 5.2 has very little effect (if any), and disable HT has also very little effect.
When examining the plot, it seems that the higher clocked settings are slightly more stable. Hinting that in a more object heavy situation the higher clocks would be slightly more favorable.
However, this might just be margin of error.