FC3 scaled to a lot of cores generally, but in situations like swimming in the middle of no where, I just had one core maxed out and the other 7 sitting idle. BF4 multi-player with a huge map and tons of players, lots of cores get used. BF4 single player, doesn't need a lot of cores. One benchmark is picking a part of a game that does scale and the other is not.įrom my experience, the most stuff going on screen, the better core scaling.
But then again you could just disable 1 module, because most games don't even come close to stressing out a 3m/6c setup)Ĭlick to expand.Just because a game can use 8 cores, doesn't mean it uses them all the time. (not exactly a gaming scenario)ġ takeaway is that you could overclock harder if you disable cores, which I think is self explanatory~ (But depending on your board's bios, there might be no reliable way to disable 1 core per module. And 1 core is nearly enough to saturate the FPU per module.Įdit: But having 2 cores per module active at the same time does not negatively impact the performance, unless both try to run at 80% load and up, or both are tasked with heavy FPU work. The test is about showing that FX 8 cores are fine as 4m/4c parts for gaming, extra cores don't add much (unless you're DX12/Mantle gaming).They add a good amount of integer compute performance (which as the video stated is nice for encoding/streaming).
The video also points out that FPS did not increase in gaming.
Core parking should be disabling those cores by default, unless they are needed, by the way.Īlso the tests in the video are about full load usage of multiple cores, also FPU limited, in some cases (which there's only 1 of per module). There shouldn't be much of a 'boost' in disabling the secondary cores.