Not only GeForce RTX 4090 and 4080, the new architecture There’s Lovelace also arrives in the field of professional video cards with NVIDIA RTX 6000a GPU per workstation which seems even more powerful than the GeForce RTX 4090, and not just because it has double the VRAM memory, well 48GB of GDDR6 with ECC.
The new video card in fact pushed by una GPU AD102 in its almost complete form, with 18176 CUDA core active against the 16384 of the flagship GeForce RTX. This allows it to also have 568 4th generation Tensor cores and 142 3rd generation RT cores available, while the memory interface remains at 384 bits.
The RTX 6000 also exhibits ben three NVENC and NVDECas well as support for decoding and encoding AV1, to better deal with the management of multimedia formats. The GeForce RTX 4090 and 4080 only have two NVENC and NVDEC. Characterized by a design dual slot with active cooling and the presence of four DisplayPort 1.4 in the rear, the RTX 6000 like the new generation GeForce equipped with a PCI Express 4.0 x16 interface and has a 16-pin CEMS auxiliary power connector.
Pending full specifications, NVIDIA declares a “maximum consumption” of 300W, well below the declared TGP of the RTX 4090 equal to 450W and lower than the 320W of the RTX 4080 16 GB. Evidently NVIDIA has set it up much lower clock ratesthan those of gaming solutions, nevertheless declares performance up to 2-4 times higher than the previous RTX A6000. Regarding availability, NVIDIA makes an appointment at Decemberwhile the price was not disclosed.