AMD Ryzen 8000 series APUs codenamed Strix Point are already in the works

AMD has charted plans to unveil the revolutionary Zen 5 architecture in 2024, spearheading the Ryzen 8000 series that also features a “big.LITTLE” architecture. Furthermore, its APU will incorporate the RDNA 3.5/3+ architecture, with the debut of these innovative products potentially eclipsing many individuals’ initial forecasts.

As per Phoronix, the AMDGPU LLVM backend/compiler has introduced GFX1150 and GFX1151, the former having been mentioned previously and the latter making its first appearance in the LLVM 17 compiler, found within Linux and other open-source codes. Following an analytical approach, the GFX1150 corresponds to the codenamed product “Strix Point”, whereas the GFX1151 aligns with “Strix Point Halo”. Evidently, AMD is adhering to its timeline and pushing the development process forward.

It has been previously reported that AMD is in the throes of developing several APUs that fall under the Ryzen 8000 series, inclusive of Strix Point and Strix Point Halo.

Strix Point, persisting with the single-chip design, introduces a mixed architecture with Zen 5+Zen 5c in the CPU section, the CU count in the GPU section expands to 16, and the TDP range peaks at 54W. Strix Point Halo, also referred to as “Sarlak”, is a high-end APU that adopts the chiplet design, with the CPU section brimming with up to 16 cores, a GPU section based on the RDNA 3.5/3+ architecture offering 40 CUs, and a TDP range stretching to 120W.

It is reported that when Strix Point’s power consumption is set at 50W, the CPU performance surpasses the existing Phoenix chip by 35%, while the GPU performance aligns with Nvidia’s RTX 3050 Max-Q. As for Strix Point Halo, its performance is far superior; with a power consumption set at 90W, the CPU performance exceeds the existing 16-core Dragon Range by 25%, while the GPU performance equates to that of Nvidia’s RTX 4070 Max-Q (90W).