Nvidia has officially unveiled the RTX A series of graphics cards during GTC Digital 2020. Consisting of the RTX A6000 workstation graphics card and the RTX A40 data centre graphics card, the new RTX A series was designed to help professionals deal with emerging enterprise workloads, “from the desktop to the data centre”. As it seems, the “Quadro” and “Tesla” naming has been dropped on the upcoming graphics cards.
Due to the COVID-19 pandemic, content is created and consumed in different ways. Teams are now more dispersed, but they keep “working remotely on high-resolution content”, using various workloads that require an enormous amount of resources. The RTX A series was created to help professionals tackle their tasks, offering all the benefits that come included with Nvidia Ampere architecture. Just like other Ampere graphics cards, the new RTX A series will feature 2nd generation RT cores, 3rd generation tensor cores, the latest CUDA cores, GDDR6 memory, and PCIe 4.0 interfaces.
Both cards will use the GA102 GPU with 84 streaming processors, totalling in 10752 CUDA cores. These graphics cards will also feature NVLink and 48GB of VRAM, but at different clock speeds: the RTX A6000 comes at 16Gbps (maximum bandwidth of 768GB/s), while the RTX A40 comes at 14.5Gbps (maximum bandwidth of 696GB/s). The RTX A40 features 3x DisplayPort 1.4, but they are disabled by default, which is expected given that it’s a data centre graphics card. On the other hand, the RTX A6000 comes with 4x DisplayPort 1.4 ready to be used. Both cards feature a TDP of 300W and a new 8-pin power connector.
KitGuru says: Do any of you regularly use professional GPUs like the Nvidia Quadro? Will your workstation be getting an Ampere upgrade?
The post Nvidia officially unveils the RTX A6000 and RTX A40 first appeared on KitGuru.