Battlefield 1 PC graphics benchmark review - AMD FX 8370 versus Intel Core i7 5960X and DX11/12 (Radeon R9 Fury)

AMD FX 8370 versus Intel Core i7 5960X on DX11/12 (Radeon R9 Fury)

Battlefield 1 is now a full DirectX 12 title. The driver overhead results as such should be much better (if properly implemented). In the past we have seen AMD perform much worse due to lower per core (IPC) performance these processors deliver. DirectX 12 can really help in this particular matter as CPU overhead can be cut off by a significant margin. As such we figured it'd be nice to compare inbetween two eigh-core processors (AMD FX 8370 versus Intel Core i7 5960X) both clocked at 4.3 GHz. 

We again take the two processors and respective platforms. Both setups wun Windows 10 64-bit all patched up, both setups run 16GB of memory. AMD has dual-channel 2133 MHz. The Intel platform is quad-channel 2400 MHz. Clearly Intel should be miles ahead here, ... right? For the rough FPS measurements have a peek at the chart above. On this page we start with the AMD Radeon R9 FURY (ASUS AIB version). The chart shows all tested resolutions versus DX11 and DX12 and the Core i7 5960X and AMD FX 8370. Again, both processors are clocked in the 4.3 GHz Turbo range, the AMD FX processor is half as fast per core and as such IF a graphics card is fully enabled in DX12, could offer better support. The long story short is that for the FURY the differences again are marginal, a couple of frames per second at 1080P and 1440P at best.

But let's zoom in and analyse with FCAT (frame capture analysis).

If we look at the FCAT percentile chart then (at WQHD) we do see a bit of a shift in FPS, the fastest processor and DX3D API is the AMD FX 8370 in combo with DirectX 11 for the the FURY. This however seems to strictly happen at 2560x1440. The differential remains enaly to close to call for the other three configurations though. But yeah, that is a marginal win. Again this however is a very tricky measurement as explosions in the game can have a dramatic negative effect on percentiles. So please do take that into account. 

And then once we we look at frametime latency, that's where you can see the bigger and more proper picture, all results are very close to each other. The big spikes are once again grenade/mortar impacts. We do see two framedrops, both occuring in DX12 modus at 24 seconds. I have noticed some people complainign about small erratic behaviour in DX12 modus, albeit we did not visually notice it ourselves. We cannot stress enough that Battlefield 1 does not require a 1200 USD processor to run. The differences in-between DX11 and 12 relative to a weaker processor remains very trivial.