Introduction

We were invited by AMD to Los Angeles last week, to be part of several technical presentations by the company’s engineering leads to understand its new Zen 5 CPU microarchitecture better. AMD Zen 5 marks the fifth generation of AMD’s winning Zen series CPU microarchitectures that have turned a once written-off company in the processor market back to competitiveness. It seems like just a couple of years have gone by, but AMD Zen is now into its seventh year, and the company has transitioned two desktop sockets, five chipset series, and seven processor lines. The new Zen 5 microarchitecture powers the new AMD Ryzen 9000 series “Granite Ridge” processors on the desktop, the all important Ryzen AI 300 series “Strix Point” processors on notebooks, and the 5th Gen EPYC “Turin” server processors. AMD is planning to launch the new server processors a little later this year, but has dedicated Summer 2024 to its client segment—desktops and notebooks.

The new Zen 5 microarchitecture builds on AMD’s proven CPU core technology, and doesn’t try anything fancy like its generational counterpart from Intel, the Lion Cove P-core used on Arrow Lake. Zen 5 still brings a double-digit percentage IPC gain over the previous generation, and introduces several efficiency improvements over Zen 4 thanks not just to its newer 4 nm process, but also a host of other innovations. AMD was able to increase clock speeds, and lower TDP across the desktop processor lineup, and still achieve good performance gains.

The notebook processor series is particularly important for AMD, as the notebook segment undergoes an upheaval in the form of native AI acceleration, the introduction of Microsoft Copilot+ standard, and the need for a processor to feature a 40 TOPS-class NPU to meet it. The new Ryzen AI 300 series mobile processor based on the Strix Point monolithic silicon does just this, combining not just up to 12 CPU cores based on Zen 5 or its derivative Zen 5c, but also a powerful iGPU based on the newer RDNA 3.5 graphics architecture that massively improves performance per Watt over RDNA 3; and a 50 TOPS NPU powered by the XDNA 2 architecture that introduces support for new efficient data formats, which exceeds the performance requirements of Copilot+ while retaining the power footprint of an NPU.

Over on the desktop side, the AMD Ryzen 9000 Granite Ridge processors continue using the Socket AM5 infrastructure of Ryzen 7000 Raphael, and are backwards compatible with all AMD 600-series chipset motherboards (with a BIOS update). The company is also taking the opportunity to introduce the new AMD 800-series chipsets, which shuffle up I/O features among the specific chipset SKUs, to provide better choice. AMD has retained its CPU core counts from the past several generations—you get these processors in 6-core, 8-core, 12-core, and 16-core variants. These aren’t hybrid or heterogeneous multicore processors, there are no E-Cores, every core is a P-Core.

AMD is planning to launch both the Ryzen 9000 desktop and Ryzen AI 300 notebook processors this Summer. Some of the first notebooks powered by the Ryzen AI 300 series should be announced by notebook OEMs as early as this week; while the Ryzen 9000 series desktop processors will be available to customers from July 31, 2024. The company hasn’t finalized prices of its desktop processors. In this article, we will provide a technical deep-dive of both processors, and the underlying Zen 5, RDNA 3.5, and XDNA 2 technologies, as well as their platforms, feature sets, and first-party performance claims by the company. We will, of course, have our reviews of the Ryzen 9000 series up before launch day.