AMD‘s upcoming Radeon RX 8000 series, featuring the new RDNA 4 architecture, has made an unexpected appearance on Geekbench, offering a glimpse into its core specs. The leaked data indicates that the GPU, identified as “GFX1201,” is a Navi 48 SKU. It features 56 compute units (CUs) and 16 GB of VRAM, confirming earlier rumors. This puts the GPU between AMD’s RX 7700 XT and RX 7800 XT in terms of compute units.
Benchmark Results and Expected Improvements
The benchmark shows the Radeon RX 8000 GPU running at a clock speed of 2.1 GHz, lower than current RDNA 3 GPUs, which typically clock around 2.5-2.6 GHz. However, these are early figures, and the final product may have higher frequencies. The GPU’s OpenCL performance appears underwhelming compared to existing models like the Radeon RX 7900 XT, scoring significantly lower in early tests. Yet, this is not surprising for a pre-production unit. As more samples are tested, better results are expected.
Further details about the memory confirm a 256-bit bus interface, aligning with other high-end RDNA 4 cards, while some models might feature a 192-bit interface and 12 GB of VRAM. Reports suggest that AMD will continue using GDDR6 memory, clocked at around 18 Gbps.
More information will likely be unveiled at CES 2025, where AMD is expected to officially announce the RDNA 4 lineup, focusing on delivering better performance and efficiency, especially in ray tracing.
