NVIDA recently unveiled its consumer-targeted GeForce RTX30 series ampere cards, and soon, the company will unveil quadro RTX graphics cards aimed at its Prosuma and workstation ampere lineup.
NVIDIA Quadro RTX (Ampere) graphics card with full fat GA102 GPU and 48 GB of GDDR6 memory
The NVIDIA Quadro RTX lineup was first introduced with Turing GPUs and will soon be upgraded to the new Ampere graphics architecture. In a tweet message Moore’s law is dead (Via videocardz), he says the image is for a next-generation Quadro RTX graphics card with a full-fat GA102 GPU.
The leaked image we see is from the NVIDA Control Panel which lists several key features of the card itself. It is not clear what the card will be called, but according to Copaito 7km, the card could probably go as high as 6000 on the Quadro RTX, which means that in the future there will be 8000 seats on the high-end Quadro RTX.
Bigger ampere cards are coming, and I have pictures … pic.twitter.com/XxQK29eQ9i
– Moore’s Law is Dead (@Mursolavisedad) September 12, 2020
A. 6000😃
– kopite7kimi (kopite7kimi) September 13, 2020
Specifications list 10752 CUDA cores equipped with 84 SMs. This is what we see in the full-fat Ampere GA102 GPU which should be faster than the GeForce RTX 3090 because it packs 2% more cores but it mostly depends on the clock speed that drives the card. The Quadro RTX A6000 operates at a boost clock of 1860 MHz which should be equivalent to about 40 TFLOP of compact horsepower.
In terms of memory, the card will have a total of 48GBDDR6 memory running across the 384-bit wide bus interface. This should be equivalent to a total bandwidth of 768 GB / sec. Now the bandwidth is less than the NVIDIA GeForce RTX 3090 which should be about 936 GB / setting using GDDR6X memory. This is certainly an interesting choice and could mean that we see next year’s GDDR6X implementation with a high-end Quadro SQ as high-density GDDR6X chips go into mass production in early 2021. A new Titan variant is also being discussed. Coming soon but we have no specifications.
NVIDA Quadro RTX Series Graphics Card Lineup:
| Graphics card | GV100 structure | RTX 5000 panel | RTX 6000 panel | RTX 8000 panel | RTX panel (amperes) |
|---|---|---|---|---|---|
| GPU | Volta GPU | Turing GPU | Turing GPU | Turing GPU | Ampere GPU |
| GPU process | 12nm | 12nm | 12nm | 12nm | 8nm |
| Size size | 815 mm | 545 mm | 754 mm | 754 mm | 628 mm |
| GPU course | 5120 cores | 3072 core | 4608 core | 4608 core | 10752 core |
| Color tensor | 640 cores | 384 cores | 576 cores | 576 cores | 656 cores |
| Clock boost | 1.62 GHz | 1.81 GHz | 1.77 GHz | 1.77 GHz | 1.86 GHz |
| Single accuracy | 16.66 TFLP | 11.2 TFLOPS | 16.31 TFLOPS | 16.31 TFLOPS | 30 TFLP |
| Ray tracking spec | N / A | 6 giga ray / sec | 10 giga rays / sec | 10 giga rays / sec | TBA |
| VRAM | 32GB HBM2 | 16GB GDDR6 | 24GB GDDR6 | 48GB GDDR6 | 48GB GDDR6 |
| NVCOR VRAM | N / A | 32GB with NV LINK | 48 GB with NB LINK | 96 GB with NV LINK | 96 GB with NV LINK |
| Memory bus | 4096-bit | 256-bit | 384-bit | 384-bit | 384-bit |
| Memory bandwidth | 870 GB / s | 448 GB / s | 672 GB / s | 672 GB / s | 768 GB / s |
| TDP | 250 W. | 160W | . 200W | 5 225W | TBA |
| Price | 000 9000 US | 00 2300 US | 00 6300 US | 00 10000 US | TBA |
| Lunch date | 2018 | Q4 2018 | Q4 2018 | Q4 2018 | Q4 2020? |
We also don’t know if NVIDA will use its GA102 or GA100 GPU based on Ampere Architect for use in its flagship Quadro and Titan graphics cards. We hear more about the graphics card’s Quadro RTX ampere line but we don’t expect the news in advance we’ll keep you posted.
Prone to fits of apathy. Music specialist. Extreme food enthusiast. Amateur problem solver.
