Skip Navigation

InGameLoop relies on readers. We may earn commissions when you purchase through our links. Check Affiliate Disclosure

AMD prioritizing multi-thread over single-thread performance for next-gen Zen 5 CPUs according to new rumours

AMD prioritizing multi-thread over single-thread performance for next-gen Zen 5 CPUs according to new rumours

Recently rumors have emerged suggesting that AMD is shifting its focus towards enhancing multi-threaded performance in its next-generation Zen 5 CPUs, potentially at the expense of single-threaded capabilities. This speculation has stirred significant interest among tech enthusiasts and industry experts alike, as it hints at a strategic pivot in AMD’s design philosophy.

The source of these rumors, the YouTube channel RedGamingTech, has provided detailed insights into AMD’s purported approach. According to the channel, AMD recognizes the substantial financial rewards tied to server chip sales, a sector that holds greater prominence than desktop CPUs for gaming enthusiasts. To this end, AMD typically employs a unified CPU core design, which is then adapted for deployment across both server and client PC platforms.

This design strategy involves the utilization of identical CPU core chipsets for both domains, supplemented by application-specific chipsets tailored to specific tasks such as IO handling and memory control.

RedGamingTech posits that this shift towards prioritizing server performance in Zen 5’s architecture could potentially narrow the gap between maximum single-core clock speeds and maximum multi or all-core clock speeds when compared to previous Zen iterations. This adjustment may have implications for single-thread frequencies, although initial assessments suggest that any potential decrease in performance is marginal.

According to RedGamingTech, early engineering samples of the Ryzen 9 8950X – believed to be associated with Zen 5 – are reportedly achieving clock speeds of 5.6GHz, a mere 100MHz below the existing Ryzen 9 7950X.

However, it is essential to note that these figures are based on engineering samples and not final silicon, leaving room for potential optimizations before the official release.

Additionally, RedGamingTech delves into Zen 5’s IPC (instructions per clock performance), which measures the efficiency of executing a single software thread per clock cycle, irrespective of clock speed. This metric is a critical determinant of a CPU’s overall performance.

Earlier information from another source, Moore’s Law is Dead, suggested a modest 10 to 15% improvement in Zen 5’s IPC based on an internal AMD slide. RedGamingTech counters this claim by contending that the slide employed a narrow measure (SpecInt) for IPC calculations. According to RedGamingTech, broader IPC gains across various integer and floating-point workloads could potentially exceed 20%.

Assuming Zen 5 achieves comparable peak clock speeds to its predecessor, Zen 4, this could translate to approximately 20% more single-thread performance. Nevertheless, the gains in multi-thread performance are expected to be more substantial due to the narrowed frequency gap between single and multi-threaded workloads.

While the emphasis on multi-thread performance may lead to a marginal reduction in single-thread performance, some analysts argue that an exclusive focus on the latter could yield even more impressive results.

In a related vein, RedGamingTech also offers intriguing insights into Strix Halo or Sarlack, a rumored mega APU that AMD is reportedly developing. This APU boasts 16 CPU cores, complemented by 40 RDNA 3.5-spec GPU compute units and a 256-bit shared memory bus – a noteworthy feature for a PC-based APU, as opposed to a console.

Comparatively, an AMD Radeon RX 6750 XT discrete desktop GPU also features 40 compute units, albeit in the older RDNA 2 specification. According to RedGamingTech, Sarlack is purportedly delivering performance on par with an RX 6750 XT paired with a 16-core Zen 5 CPU, all while operating within a power range of 75 to 125W. This achievement is particularly remarkable considering that AMD’s highest-performing 16-core desktop CPUs alone consume 170W.

As with any information based on rumors, it is crucial to approach these insights with a degree of caution. Nevertheless, as the launch window for new CPUs and GPUs draws nearer, the accuracy of these reports tends to improve, rendering this latest rumor dump both intriguing and potentially indicative of AMD’s future design direction.

LGA1700 Coolers Are Compatible With Next-Gen Intel Arrow Lake CPUs

AMD’s next gen Threadripper Pro 7000-series chips come with up to 96 cores and I long for consumer version

Nvidia’s Next-Gen Blackwell GPUs Rumored to Use Multi-Chiplet Design

Usman Haider

Journalist

Usman Haider, an acclaimed content writer and CEO of Epha Tech, offering expert digital marketing solutions. Harnessing his skills in content creation and marketing strategy, Usman sets new standards in the digital world. Having six years of experience in tech industry.