AMD is well-known for its innovation in the CPU market, and its upcoming next-gen Zen CPUs, the Ryzen 8000 series, are highly anticipated. While the release is scheduled for next year, there are already numerous rumors circulating about the new chips. One recent claim comes from the YouTube channel RedGamingTech, stating that AMD is prioritizing multi-threaded performance over single-threaded. This decision may seem surprising at first, but upon closer examination, it becomes clear that it aligns with AMD’s focus on server chips, which are their primary revenue source.
RedGamingTech argues that since server chips offer greater financial rewards for AMD compared to desktop CPUs, Zen 5’s architecture has been slightly tilted in favor of server performance. This alignment with server performance reflects AMD’s strategy of using a single CPU core design across both server and client PC platforms, with additional application-specific chiplets for each platform. These chiplets provide IO, memory controllers, and other features tailored to the specific platform’s requirements.
As a consequence of this focus on server performance, the gap between the maximum single-core clock speed and maximum multi or all-core clock speed has been narrowed for Zen 5 compared to previous Zen chips. While this narrowing may imply a potential decrease in single-thread frequencies, the impact seems minimal. RedGamingTech claims that “late engineering samples” of the Ryzen 9 8950X, said to be part of the Zen 5 lineup, are reaching clock speeds of 5.6GHz. This figure is only 100MHz lower than the existing Ryzen 9 7950X. It’s important to note that these are engineering samples, and final silicon may offer further improvements.
RedGamingTech also provides insight into Zen 5’s IPC (instructions per clock) performance, which is an essential measure of a single software thread’s performance per clock cycle regardless of clock speed. Previous information from Moore’s Law is Dead suggested a 10 to 15% improvement in IPC for Zen 5. However, RedGamingTech contests this claim, stating that the slide used a narrow measure of SpecInt to derive those numbers. Instead, they argue that the broader IPC gains across multiple integer and floating point workloads will likely be 20% or more.
Considering similar peak clock speeds to Zen 4, the projected 20% increase in single-thread performance is quite promising. Moreover, with the gap between single and multi-threaded loads narrowed, the multi-threaded performance gains are expected to be significant. Thus, it can be argued that the emphasis on multi-threaded performance leads to a substantial boost in this area, potentially outweighing any marginal decline in single-thread performance that may have occurred.
Moving beyond the Ryzen 8000 series, RedGamingTech also touches on the rumored Strix Halo or Sarlack APU. This APU is speculated to feature a 16-core CPU and 40 RDNA 3.5-spec GPU compute units, along with a 256-bit shared memory bus – a unique feature for a PC APU. To put it into perspective, an AMD Radeon RX 6750 XT discrete desktop GPU also offers 40 compute units, albeit in an older RDNA 2 spec.
According to RedGamingTech, Sarlack is said to deliver comparable performance to an RX 6750 XT combined with a 16-core Zen 5 CPU while operating within a power range of 75 to 125W. This achievement is impressive, considering that AMD’s highest-performing 16-core desktop CPUs alone consume 170W.
The rumors surrounding AMD’s Ryzen 8000 series and the focus on multi-threaded performance rather than single-threaded demonstrate the company’s strategic alignment with its server chip business. While the narrowing of the gap between clock speeds may have resulted in a minor decrease in single-thread frequencies, it appears that the impact is negligible. With projected IPC improvements and overall performance gains, the Zen 5 architecture promises to deliver significant enhancements for both server and client PC platforms.
Additionally, the rumored Strix Halo or Sarlack APU hints at AMD’s ongoing commitment to pushing the boundaries of integrated graphics performance. While all these details should be taken with a grain of salt until official announcements are made, the excitement and anticipation surrounding AMD’s next-gen processors continue to grow.