In continuation,
much of the misinformation comes from for want of a better term, fanboys I’ve seen people find a reason why every product is wrong, simply because it isn’t their companies. I’m sure everyone here is familiar with similar posts both for and against apple products.
However the blunt and honest truth is - a processor is a processor, a GPU is a GPU - and as long as it works for you, it’s fine.
I personally don’t take into account optimizations or instruction sets, as they are subject to fall out of favor, or change beyond the original specification.
A good example of this is Mantle, AMD’s answer to DirectX/OpenGL - while some parts where incorporated into Vulcan, it’s dead. Even though it was heralded as the next big thing and showed some promise.
All that matters in my opinion is the baseline FLoating Point Operation (FLOP) counts - and now it’s time to get UE4 specific.
If you plan on doing anything particularly serious with UE4, it’s very likely you will want to build the engine from source at some point - most likely using MSVC. Microsoft are surprisingly pretty ■■■■ good at staying out of hardware-specific wars, and on the MSVC tool-chain this shows - compile times per FLOP are pretty much identical on both hardware platforms [my google-fu failed, citation needed].
For both editor-use and compiling, more cores are king - thus AMD has a discreet advantage - however there’s a couple more things to consider.
Single core performance has pretty much flat-lined along with clock-speed - and that isn’t going to change unless there is some dramatic break-though in technologies.
[Source][2]
As you might expect, this leads to more cores being added - the simple reason we don’t see 32-core processors on the market is - software hasn’t caught up, programmers haven’t adapted yet.