NVIDIA Announces RTX A6000 48 GB Professional Graphics Card Accelerators
NVIDIA today introduced their RTX A6000 series of graphics playing cards, meant to conduct as graphics accelerators for qualified workloads. And the announcement marks a massive departure for the company’s advertising, as the Quadro moniker has seemingly been dropped. The RTX A6000 consists of all raytracing resources also current on client RTX graphics playing cards, and marks a solution segmentation from the company’s datacenter-geared A40. The RTXA6000 features a whole-blown GA102 chip – indicating 10752 CUDA cores powering solitary-precision compute efficiency of up to 38.seven TFLOPs (3.1 TLFOPs higher than that of the GeForce RTX 3090). Besides providing NVIDIA’s experienced driver aid and attributes, the RTX A6000 functions forty eight GB of GDDR6 (be aware the absence of the X) memory – ensuring every thing and the kitchen area sink can be saved in the cards’ VRAM. GDDR6X will not at present present the for each-chip density of GDDR6 answer, that’s why why NVIDIA opted for the lessen-accomplishing, still denser memory variant.
The RTX A6000 functions a common blower-type cooler, and presents a new lower-profile NVLink bridge that allows two of them to get the job done in tandem in just the similar system. NVIDIA vGPU virtualization technologies are supported as nicely show outputs are taken treatment of by 4x DisplayPort connectors, marking the absence of HDMI solutions. The card is at the moment mentioned for preorder at a cool and collected $5,five hundred, but with insufficient silicon to supply even to its maximum-margin datacenter customers, it continues to be to be found accurately how offered these will be in the market.