Conclusion

Having tested ArmA 3 with CPU and memory stock as well as overclocked using nine different graphics cards and three different resolutions at ultra-settings, we believe in having a thorough basis for understanding how ArmA 3 responds to changes in clock speeds. The different three different resolutions with nine different graphics cards from AMD and NVIDIA in combination with an Intel Core i7-6700K at four different presets we believe to have a thorough understanding to what this games engine responds.
A closer look at the results shows that the engine is not capable of fully loading the GPUs at 1080p as well as 1440p. Even at 2160p the GPU is only partially bottlenecking performance.

Let's first have a look at the influence different graphics cards have, talking about the extremes. We've measured that when using an AMD Fury Nano graphics card, while overclocking your CPU as well as your memory, average performance benefit, considering the three resolutions we've tested, is about 20.9 percent. Doing exactly the same but using an NVIDIA GTX 970 only results in a 6 percent increase in performance. Therefore it's very interesting to see that the frame rates in ArmA 3 show better positive response, when using AMD graphics cards. Especially the new AMD Fury cards appear to allow for better CPU and memory performance scaling, while with NVIDIA cards the GTX 980 as well as the 980 Ti are about on par in that matter.

At this point we're going to have at performance with the different resolutions. In theory we should see the largest performance benefits by overclocking CPU and memory when testing 1080p. The second most influence should be notice in the case of 1440p and 2160p should benefit the least. The reason for this is that at 1080p the load on the GPU is the lowest, thus the GPU is not bottlenecking performance. At 2160p the GPU could be a possible bottleneck but it doesn't have to be, meaning that the CPU could have an influence but also couldn't.

Checking the test results we see that this theory is spot on when it comes to 1080p as well as 1440p testing. Depending on the graphics card you're using, overclocking your CPU and memory allows for performance benefits of 21.4 percent (AMD Radeon R9 Fury) and 8.3 percent (NVIDIA GTX 970) at 1080p. Having a look at 1440p we see that frame rates can be bumped by between 21.4 percent and 9.7 percent, again depending on the GPU used. 2160p is where the results are particularly interesting, since AMD cards are not bottlenecking performance but NVIDIA cards on the other hand are. AMD cards can allow for up to 19.4 percent higher frame rates when overclocking the CPU and the memory but when using an NVIDIA card the highest increase we measured is with the GTX 980 and it's a mere 3.57 percent.

Bottom line it can be said, that ArmA 3 responds pretty well to a CPUs clock speeds especially at 1080p and 1440p. In the case of 2160p overclocking only gives you an advantage if you're using an AMD graphics card, while NVIDIA cards bottleneck frame rates. If you want to play ArmA 3 at the highest possible frame rates, using a single card, we recommend an AMD Fury based graphics card, while overclocking your CPU and memory to the maximum they do stable.