3 Commando Brigade

Apollo hop on TS I'll show you a good deal where I got mine

Sent from my Samsung Galaxy S7 Edge

If there's a website link throw it in here bud I've been looking into an upgrade too

Sitting here with my old GTX 680 which has served me well throughout the years, also on an i7 4770k at 4.2GHz and as others have said above the CPU is the most important thing when it comes to running Arma, always has been. Game runs great with anywhere from 45-55 FPS throughout on high-very high settings but of course Arma being Arma, I usually have to restart after a while otherwise my frames are shot to pieces. 

It's a limit to DX11 and the engine. Extra cores are not used to their full potential, mostly sitting at 30-35%. That's what I was referring to regards to core count not being that effective. It's the strength of the first core and its clock speed that makes all the difference rather than loads of cores at a lower clock.

Dwayne, those settings are legacy developer settings. I.e for testing the game on different hardware. All of the parameters are auto configured on launch, you're just slowing your launcher down by manually telling it what it wants to do anyway


Sent from my iPhone using Tapatalk

I've not seen a difference when playing with them settings and now I know why. Cheers Apollo

No because of the limits of the engine and directX as I said earlier.

It's just one of those internet rumours thing... put it on a website and it must be true.. right ?

Sj

@Apollo an i5 with 970 will be fine.  You'd probably not see much benefit from an 980

Remember to look at the date of "tweaks" on the net.  A lot are now legacy only, as the A3 game engine has changed a lot over the past few years.

There are two main reasons why A3 does not improve in fps with extra cores.  Firstly the main engine is largely single threaded, meaning that it has been written to only run on a single core.  Windows will move it around from core to core, but it'll only be running on a single core at any one time.  The other is due to DX.  All versions of DX prior to 12, can only communicate with the GPU from a single thread, which runs on a single core. So its a major bottleneck. 

Processor frequency is only a good comparative measure of performance when comparing very similar processors.  Just because one processor runs at a slower freqency than another from a different family, does not mean its slower.  Caches and other processor logic (e.g. branch prediction) play a huge part in the speed of a application.