Leaked NVIDIA RTX A6000 benchmark result points to uninspiring performance with only an 11% uplift over the RTX 6000
It has only been a few weeks since NVIDIA announced the RTX A6000, one of its GPUs designed for content creators, data centres and engineers. Based on NVIDIA's Ampere architecture, an early SPECviewperf 2020 benchmark result suggests that the RTX A6000 offers only minor performance improvements over the RTX 6000.
Published by AdoredTV, the RTX A6000 averaged only 11% higher scores than the RTX 6000 across eight benchmarks. AdoredTV notes that the RTX A6000 could be underperforming, or that SPECviewperf 2020 may not be representative of the card's performance. Nonetheless, it seems that the RTX A6000 does not offer the leap in performance that one may expect Ampere to offer over Turing.
NVIDIA has equipped the RTX A6000 with 48 GB of GDDR6 VRAM, rather than the GDDR6X that it uses for consumer Ampere cards. The RTX A6000 also offers 10,752 CUDA cores across its GA102 GPU, along with four DisplayPort 1.4 ports. The card supports NVIDIA NVLink too and has a rated maximum power consumption of 300 W. /