The RTX A6000 graphics card, aimed at workstation use by professional users in the field of digital information visualization, is the only graphics accelerator based on the maximum GPU GA102, which uses the full set of 10,752 CUDA cores. For comparison, the modification of the GA102-300 chip, on the basis of which the flagship GeForce RTX 3090 gaming video card is built, has only 10,496 CUDA cores on board. Thanks to the difference in the number of computing units, the net performance of the RTX A6000 is 38.7 TFops, which is 3.1 TFlops more than the gaming flagship of the Ampere series. An additional difference between the RTX A6000 and a gaming graphics card is that it is equipped with 48GB of GDDR6 video memory with a 384-bit bus.

The RTX A6000 comes with four DisplayPort 1.4 connectors, with no HDMI connectors. The cooling of the RTX A6000 components is handled not only by the massive heatsink but also by the tangential fan. The card supports the NVLink interface, which allows combining the processing power of two graphics accelerators.

The announced technologies include support for GPU virtualization through tools such as NVIDIA GRID, NVIDIA Quadro Virtual Data Center Workstation, and NVIDIA Virtual Compute Server.

Resource VideoCardz reports that PNY’s RTX A6000 has appeared in the assortment of a large American retailer B&H, which has estimated it at $ 5500. The seller offers to pre-order and informs that the card will be delivered within 2-4 weeks. According to the source, the offer is very interesting, since the predecessor in the face of the Quadro RTX 6000 was originally sold for $ 6,300, but at the same time, it offered only 24 GB of video memory.