NVIDIA released the RTX A400 and RTX A1000 models, adding to its lineup of professional workstation GPUs. Based on the Ampere architecture, these entry-level cards are the first in NVIDIA’s array of low-end desktop professional graphics cards to come with Tensor Core and RT Core technology.
Based on Samsung’s 8nm GA107 core, the RTX A400 and RTX A1000 are intended to run within a 50W thermal design power (TDP). They are built using half-height, half-length (HHHL) form factors and active cooling systems. Each board has four Mini DisplayPort 1.4a outputs.
The 2021-released T1000 and T400 GPUs, among the final NVIDIA professional GPUs without the RTX moniker, are expected to be replaced by these new models. More than three times the generative AI processing speed of its predecessors, the RTX A1000 offers a notable performance gain. Its RT Core also improves the effectiveness of apps in professional fields like video editing and CAD.
The specifications for the RTX A400 and RTX A1000 are as follows:
RTX A1000: 2304 CUDA cores, 72 third-generation Tensor Cores, 18 second-generation RT Cores, a boost clock of 1460 MHz, 8GB of 12Gbps GDDR6 memory across a 128-bit interface, along with 2 NVDEC decoders and 1 NVENC encoder.
RTX A400: 768 CUDA cores, 24 third-generation Tensor Cores, six second-generation RT Cores, a boost clock of 1755 MHz, 4GB of 12Gbps GDDR6 memory across a 64-bit interface, with 1 NVDEC decoder and 1 NVENC encoder.
The RTX A1000 is already available through NVIDIA’s global distribution network, while the RTX A400 is expected to hit the market in May, with broader availability from manufacturers by the summer.