A day after Nvidia announced the RTX 4000 series of gaming cards, it also showed off the next generation of workstation cards based on the same architecture, Ada. The company has not disclosed how much it would cost so far.
We have to stress workstation GPUs here, these are meant for professionals and will command hair-raising price points: The RTX 6000 is listed on several US retailers with prices ranging from $7,378 to $8,210, according to data compiled by VideoCardz.
Nvidia RTX 6000 (Ada) on CompSource and ShopBLT pricing
The Nvidia RTX 6000 card has 18,176 CUDA cores, which is 11% more than the RTX 4090’s 16,384 cores. However, it appears that it will run at a lower power: the TDP is only 300W compared to the gaming card’s 450W. Even so, the maximum boost clock is capped at the same 2.5GHz. Another thing to note is that cooling has switched to a blower-type fan, which is the preferred solution for professional environments (it’s noisier, but has less chance of overheating).
The main difference between the RTX 6000 and RTX 4090 is graphics memory. The workstation GPU comes with 48GB, double what the game card has. It’s also a different type of memory, GDDR6 with ECC (which is built-in error correction). The memory bus remains 384 bits wide, so bandwidth is basically the same (21Gbps vs 20Gbps).
| GPU functionality | NVIDIA RTX 6000 |
| GPU memory | 48GB GDDR6 with error correcting code (ECC) |
| Show ports | 4 DisplayPort 1.4 ports* |
| Maximum power consumption | 300W |
| Graphic bus | PCIe Gen4x16 |
| Form factor | 4.4″ (H) x 10.5″ (L) dual slot. |
| Thermal | Active |
| vGPU software support | NVIDIA vPC/vApp, NVIDIA RTX virtual workstation |
| Ready for VR | Yes |
Note that Nvidia no longer uses the Quadro branding for these workstation GPUs. Even more confusing is that the card the RTX 6000 (Ada) is replacing is called the RTX A6000 (Ampere).
Start a new Thread