Architecture

At the heart of the Ryzen 3 2200G is the new 14 nm "Raven Ridge" silicon. This chip is a full-fledged SoC which combines a CPU, an integrated GPU (iGPU), a dual-channel DDR4 memory controller, platform I/O, and integrated southbridge. Unlike the 8-core "Summit Ridge" silicon AMD launched the "Zen" CPU microarchitecture with, "Raven Ridge" has just four "Zen" CPU cores, but a rather large iGPU based on the "Vega" graphics architecture.

Before its debut late-2017 in its mobile avatar, we were expecting AMD to increase CPU core-counts for its first "Zen" based APU. Unfortunately, that wasn't to be, but the quad-core CPU you do get has SMT (multi-threading) enabled on the Ryzen 3 2200G. Why four cores? This is partially because of the way AMD designed multi-core chips using its "Zen" micro-architecture.

AMD decided that four "Zen" CPU cores group better for its product managers to carve out SKUs with when designing the CCX (CPU complex). Each CCX has four CPU cores, each with dedicated L2 caches and a shared L3 cache. "Summit Ridge" has two CCX units running in a 2+2 configuration; while "Raven Ridge" combines one CCX (4+0) with the Vega 11 iGPU.



It's important to note that the CCX of "Raven Ridge" is slightly different than the one found on "Summit Ridge" because its shared L3 cache is halved to just 4 MB. This is also why you can't directly compare CPU performance of the Ryzen 3 2200G clock-for-clock with the Ryzen 3 1200 "Summit Ridge," which has two cores enabled per CCX and 4 MB of L3 cache per CCX. It's the "4+0 vs. 2+2" face-off we wanted, though 4+0 doesn't have the same amount of cache as 2+2. Each of the four cores still has 512 KB of dedicated L2 cache. The 2400G features SMT, while the 2200G lacks it.

Another major difference between "Summit Ridge" and "Raven Ridge" is the latter's smaller PCI-Express (PCIe) root complex. While "Summit Ridge" features 16 PCIe gen 3.0 lanes toward PEG (PCI-Express Graphics), in addition to 4 lanes allocated to the chipset-bus and another 4 lanes driving an M.2 NVMe slot, "Raven Ridge" puts out just 8 lanes toward PEG, besides 4 lanes as chipset-bus and 4 lanes for the M.2 NVMe slot. What this means is that the second PCI-Express 3.0 x16 slot on X370 chipset motherboards won't work, and whatever graphics card you install on the first slot will run at PCI-Express 3.0 x8 (half bandwidth). This design choice is bewildering due to the potential PR drama. The actual performance lost from PCIe x8 3.0 vs PCIe x16 3.0 is minimal. The Vega integrated graphics is connected to the CPU using Infinity Fabric.

The iGPU implements AMD's latest "Vega" architecture, with separate regions on the "Raven Ridge" silicon for the iGPU's key components and the display controllers with multimedia processing. It is endowed with up to 704 stream processors spread across 11 compute units, 44 TMUs, and 16 ROPs. It features DirectX 12 and Vulkan support and is capable of hardware-accelerated decoding of HEVC/x265/VP10. It has the necessary DRM features to enable playback of Netflix 4K-HDR and other emerging high-resolution content services. Its display controllers support 4K Ultra HD and 5K displays through DisplayPort 1.4.

The Zen Architecture

The oldest reports about AMD working on the "Zen" architecture date back to 2012, when AMD re-hired CPU core designer Jim Keller credited with the original winning K8 and K9 architecture designs to work on a new core architecture to succeed "Bulldozer." AMD continued to invest in the "Bulldozer" IP in the form of incremental core updates, hoping that trends in the software industry towards parallelization would improve, giving it a big break in price/performance. Those trends, in the form of DirectX 12 and Vulkan 3D APIs being multi-core friendly, came in a tad late (towards late 2016). Four years of work by a team dedicated to its development, led by Jim Keller, resulted in the "Zen" core.



At the heart of the "Zen" core are two very important innovations - a very "intelligent" branch-prediction system that uses neural nets (yes, of the same kind that power deep-learning machinery) to predict branches in code and load the most appropriate instructions and allocation of core resources and a 1.5X increase in issue width and execution resources, besides a 1.75X increase in the instruction scheduler window. Intel had been beating AMD in core performance and efficiency in exactly these two areas, and AMD finally addressed it instead of throwing in many more hardware resources without addressing the branch-prediction issues. "Zen" also features an up-to-date ISA instruction set including AVX2, FMA3, and SHA.


Interestingly, AMD talks about refinements to the micro-architecture itself in its Ryzen 2000G press deck for these processors, which speak of improvements to the various SenseMI components.

The AM4 Platform

What sets "Summit Ridge" apart from Intel dies, such as "Kaby Lake," is that it is a full-fledged SoC (system-on-chip). It integrates both the northbridge and southbridge. In addition to memory and PCIe, socket AM4 processors also put out USB 3.0 and two SATA 6 Gb/s ports. The platform still has something called a "chipset," but it only serves to increase connectivity options, such as adding more SATA ports, USB 3.1 ports, and a few more general-purpose PCIe lanes. On Intel's platforms, the PCH (platform controller hub) serves the functions of the southbridge, while the northbridge is fully integrated with the processor.



AMD has three chipsets for Ryzen - the X370 for high-end desktops, which supports proprietary multi-GPU technologies, such as NVIDIA SLI, the mid-tier B350 chipset with a slimmer connectivity feature set, and the entry-level A320 chipset for low-cost desktops (which lacks multiplier overclocking support). NVIDIA SLI will not work on X370 chipset motherboards if paired with "Raven Ridge" processors due to the PCIe lane limitation of the processor.