Yeah, it is "Basically" to an Extent, 2x 4 Core Chips on a Single package Linked through an Underlying Layer.

-Everything is Fab'd in 4 Core/8 Thread CCX's,
-Those are put together on the "AMD Infinity Fabric" with a Shared Cache in the Middle,
-Which allows the CCX's to Communicate w/ Each Other, or the CCX to Comunicate w/ the GPU Block.
-The Blocks and the Infinity Fabric are then placed on the Substrate Package

The Part Reddit is hung up on is "The CCX's", Since they are Separate, everyone is saying they have Dual 4c CPUs instead, which is not entirely the case.

Windows Scheduler for one, doesnt know, do not take a task from Thread 2 and move to thread 7 in a single thread benchmark, or make threads 3 and 7 work on the same task

Windows Scheduler Knows how to distribute work on the Intel Chip Threads already, as they used the Intel Compiler.

Linux Scheduler was Updated on Day One for Ryzen, and does not have the Problem

Windows 7 Has less of a performance hit than Windows 10.;

I'd just use Affinity mask in Windows, and Keep my Games on Cores 4,5,6,7. which could solve the problem.

Last edited by SkateZilla; 03/09/17 07:24 PM.

HAF922, Corsair RM850, ASRock Fata1ity 990FX Pro,
Modified Corsair H100, AMD FX8350 @ 5.31GHz, 16GB G.SKILL@DDR2133,
2x R7970 Lightnings, +1 HD7950 @ 1.1/6.0GHz, Creative XFi Fata1ity Platinum Champ.,
3x ASUS VS248HP + Hanns�G HZ201HPB + Acer AL2002 (5760x1080+1600x900+1680x1050), Oculus Rift CV
CH Fighterstick, Pro Throt., Pro Pedals, TM Warthog & MFDs, Fanatec CSR Wheel/Shifter, Elite Pedals
Intensity Pro 10-Bit, TrackIR 4 Pro, WD Black 1.5TB, WD Black 640GB, Samsung 850 500GB, My Book 4TB