A next-gen RTX Quadro has more CUDA cores and memory than the 3090: new leak
A popular graphics-tech tipster has shared what he claims are specs for an upcoming Ampere-based RTX Quadro card. These pro-grade graphics may be popular indeed among content creators and designers should the parameters in the new leak prove accurate.
According to Moore's Law is Dead, the second-gen RTX Quadro will rock 10572 CUDA cores, which is even more than those of the recently-launched flagship gaming card of the same new architecture (10496). The former GPU will also have 48GB of memory: double that of the 3090.
Then again, the latter has GDDR6X memory, whereas the "new Quadro" apparently does not. Furthermore, its memory speed is given as 16.0 gigabits per second (Gb/s), which, while extremely fast, does not beat the RTX 3090's 19.5Gb/s spec.
The possible Ampere-based Quadro also has a 384-bit memory bus, and, thus, a bandwidth of 768 gigabytes per second (GB/s). Again, this fails to beat the 3090. Then again, the rumored card may not be directed at gaming and thus may be more effective in tasks that do not prioritize the RTX 3000-series' flagship's advantages.
In addition, the putative upcoming card does destroy the current highest-end Quadro RTX 8000 in all of these aspects. That brings us to the subject of its possible successor's name. Some correspondents speculate that it will follow an Ax000 convention, whereas Videocardz proposes it will simply be the RTX 8100 on its launch. This is currently thought to take place during NVIDIA's next virtual event, scheduled for October 5, 2020.