In 2018, when presenting Turing architecture, NVIDIA first announced professional Quadro RTX cards – Quadro RTX 8000, RTX 6000 and RTX 5000 and RTX 4000, and only later did GeForce cards appear for gamers. This year, with the premiere of the Ampere architecture, the situation turned around, because the manufacturer first attacked with graphics cards for players. Only at the beginning of October, the manufacturer discussed the first details about professional Quadro cards, based on the Ampere architecture. The first details of the Quadro RTX A 6000 and Quadro RTX A 40. The first one was prepared for workstations, while the second one was designed for data centers. While Quadro RTX A 6000 has single-turbine active cooling, Quadro RTX A 40 has a passive cooling system. The company finally gave more details on the availability of the RTX A model 6000 – by the way, it was decided to drop the name Quadro.

NVIDIA announced the possibility of ordering the professional RTX A chip 6000 based on the Ampere architecture. The price of the new card is 4650 dollars, so less compared to the Quadro RTX 8000.

NVIDIA RTX A 6000 is based on the full Ampere GA core 102 with 10752 CUDA cores, 336 3rd generation Tensor cores and 84 RT units for hardware ray tracing acceleration. The novelty is equipped with 48 GB of GDDR6 memory (with support for ECC error correction), similar to the previous generation model – Quadro RTX 8000. This time, the effective clock speed 16000 MHz was used, which translates into a bandwidth of 768 GB / sec. NVIDIA RTX A card 6000 uses 384 – bit memory bus. There is also a single, 8-pin power connector and the well-known turbine-based cooling system. The entire card has a TDP of 300 W, slightly less than the GeForce RTX 3080 / RTX 3090. The core clock is lower, but the manufacturer did not provide a specific value. The declared capacity FP 32 is 38, 7 TFLOPS; RT Core – 75, 6 TFLOPS (by leaps and bounds), while the power of Tensor Core is a maximum of 309, 8 TFLOPS. The new chip has 4 DisplayPort 1.4a connectors and supports the PCIe 4.0 interface.

Engineer amazing products, design state-of-the art buildings, drive scientific breakthroughs, and create

immersive entertainment with the world’s most powerful graphics solution.

The #NVIDIARTX A 6000 is now available.

– NVIDIA Design (@NVIDIADesign) December 15, 2020

NVIDIA RTX A 6000 supports connecting cards using the proprietary bridge NVLink. In this case, however, we can connect up to two RTX A systems 6000, thus gaining access to 96 GB of memory. The manufacturer has officially announced the start of sales of a professional card for workstations. The question of price is quite interesting. The NVIDIA RTX A system 6000 was priced at 4650 dollars (approx. 40 thousands of zlotys excluding VAT). For comparison Quadro RTX 6000 and Quadro RTX 8000 were valued at 4000 and 5500 dollars. Despite the higher processing power, RTX A 6000 is therefore cheaper than Quadro RTX 8000. As mentioned, NVIDIA finally abandoned the “Quadro” name, aligning the name of its accelerators to the common “NVIDIA RTX” family. At the same time, it was announced that the drivers were changed for the current Quadro RTX systems – from now on they are called NVIDIA RTX Enterprise Drivers.

Source: NVIDIA, VideoCardz