Radeon 880M struggles to surpass its predecessor in OpenCL test, despite RDNA 3.5 architecture
Early benchmarks show AMD's new integrated GPU falling short of expectations, but slow system memory and potential power throttling might be to blame.
While AMD's upcoming Strix Point APUs with RDNA 3.5 integrated graphics have generated excitement, recent OpenCL tests of the Radeon 880M are raising eyebrows for the wrong reasons. Despite boasting a new architecture and higher clock speeds, the 880M only managed a score of 30,521 in Geekbench's OpenCL benchmark.
This score falls short of even the fastest Radeon 780M results, which achieved a score of 31,337, making the new iGPU effectively on par with its predecessor. It also places the 880M in the same performance bracket as Intel's 8-core Arc iGPU.
This disappointing performance directly contradicts ASUS's claim of a 15% performance uplift for the 880M over the 780M in 3DMark Time Spy. However, several factors could explain this discrepancy.
The tested Zenbook S16, equipped with the Ryzen AI 9 365 APU and Radeon 880M, utilized relatively slow 1,868 MT/s DDR5 memory. Since iGPUs rely on shared system memory, this could significantly bottleneck performance.
Additionally, the laptop's power settings during testing remain unknown. It's possible the Ryzen AI 9 365 operated under a restricted TDP, limiting the 880M's potential.
Therefore, it's premature to draw definitive conclusions about the Radeon 880M's performance based solely on this OpenCL score. Factors like system memory speed and power limits can significantly influence benchmark results. Awaiting comprehensive reviews that consider these variables will provide a more accurate picture of the Strix Point APUs' true capabilities.
