NVIDIA has just announced its new professional GPU, the RTX A6000. The A6000 is the new-age professional GPU, the first one based on the Ampere family of GPUs. We have previously reported about the company’s Quadro GPU lineup, but it seems like NVIDIA is going away from that branding with the new generation.
The RTX A6000 is based on the GA102 GPU, which is the same GPU as seen in the GeForce RTX 3080, but with a different configuration. AnandTech reports the CUDA core count at 10752, with 336 Tensor cores. It comes with 48 GBs of GDDR6 memory in a 384-bit bus, with ECC support. The card is rated for 300W of power via a 4+4 pin EPS connector.
You also get NVLink support, as well as support for the whole suite of NVIDIA’s GPU virtualization software. The card will be available in early 2021, in OEM systems.
The NVIDIA A40 is another card that was launched alongside the A6000. It’s very similar on-paper (minus the reduced GDDRS6 spec speed), except that the A40 is designed to be passively cooled. The form-factor is designed to be used in high-density servers that have dedicated cooling solutions. The card will be available in early 2021, in servers.
What are your thoughts? Talk to us in the comments, down below!