With the release of the Ada Lovelace architecture GPU, Nvidia not only updated the GeForce graphics card product line in September this year but also launched the first workstation graphics card based on the new architecture, NVIDIA RTX 6000 Ada Generation, which open a new era of neurographics that combines AI and simulation.
Although NVIDIA RTX 6000 Ada Generation is also equipped with AD102 GPU, its specifications are better than GeForce RTX 4090, and only 2 sets of SM are disabled. There are 142 sets of SMs, 18176 CUDA cores, about 11% more than the latter, and the boost clock is 2.5 GHz. It is equipped with 48GB of GDDR6 memory with ECC, the memory interface width is 384 bits, the rate is 20 Gbps, and the memory bandwidth is 960 GB/s.
The thickness of this graphics card is a double slot, the power consumption is 300W, and the requirements for heat dissipation are much lower. The radiator also adopts a relatively traditional turbo cooling design, and the graphics card does not support the NVLink bridge. In terms of display output, there are 4 DisplayPort 1.4 interfaces, and no HDMI interface is provided.
NVIDIA claims that NVIDIA RTX 6000 Ada Generation will have a two-to-four-fold performance increase in enterprise environments, and designers and engineers can drive cutting-edge simulation-based workflows to build and verify more complex designs; artists can create more engaging content and build immersive virtual environments; scientists, researchers, and medical professionals can have supercomputing power on their workstations to accelerate the development of drugs and procedures.
RTX 6000 Ada Generation is currently on the market, and the price offered by retailers is $7350. NVIDIA has already published an official driver for RTX 6000 Ada.