Ryzen and emulation.

[TheLuciferPhenotype]

Wouldn't the a.i. branch predictor at least to some degree enhance processing of bizarre workloads, say emulation? Remember kids, the OG PS4 was special silicone in that it had a bus from CPU to GPU, being up and cozy cutting the latency further as is relevant to out of order processors. The shared floating point and cache resources was a bit like running with ones shoelaces tied together on the old Jag so its an understandable flip from supercomputings fledgling daughter the CPU that thought it was a GPU in an odd case where the radeon compute shaders are doing x86 tasks. Ryzen shares some attributes that are decidedly the future of all computing. Efficiency per watt.

The airforce used the PS3 as it scales system to system performance per watt (as well as cost) excellently thanks to the infamous 1:1 ratio, which cleverly, ryzen adopts where it can. Im not saying its Cell, but im saying a 9 thread retro redemoing at higher resolution with minimal input might be a thing.  

[TheLuciferPhenotype]

Worded perhaps more penetratingly: PS5 may in fact be very much capable of making the metal gear solid 4 experience not a 15fps chugfest. Modern 240hz TV voodoo can somewhat hide this laughable output rate at least. Last update I checked the Xbox One X, ceased support for 4k oldy titles. Particularly the OG Xbox line. As a collector of hardware, I took this as a slap in the face. Hope this is retained if they indeed go this route. 

I wonder if we can get an adoption outside consolespace as well. In other news, how come no one is talking about AMD's DX11 drivers being single threaded? I mean has this changed ever? What about freesync functioning under DX9? See, in my tinkermanship with this here ancient now i7 4720HQ from Asus, I get the idea G sync modules, well...… aren't actually onboard whatsoever. Yet nvidias drivers still support freesync, though perhaps, mislabeled. AMD if it hasn't already, should consider changing this. They are the only reason I haven't given them a chance since GCN 2, which honestly overall compared to its launch competition, was the superior architecture. Asynchronous compute, words used in multiple instances of marketing, as they relate to the shader DX 12.0 feature, has since gone underground in explanation despite it being a huge deal. I mean we have turings dual math per pass, pascals "asynchronous compute" but load balancing preemption cache + logic based serial acceleration method that doesn't end up penalizing DX11 performance like AMDs architecture does as its optimized for low level API's. Shamefully, the development cost increase associated with junking hardware abstraction has made the real world and proper utilization of low level API's outside fixed hardware next to nonexistent. Hopefully ray tracing and photo texture conversion can trim some of that budgetary black hole and allow room for proper low level GPU utilization, perhaps handled by game engine studios and A.I. cheapening the process. Id really like some techy sort of outfit to cover these topics in more detail. Digital Foundry failed me, but of course im joe blow from the internet and my choice of screaming wasn't loud enough I guess.