During the last few months, getting an RTX 3000 at reasonable prices was almost impossible, but now that things are easing, we have a curious situation: everything points to NVIDIA will present its RTX 4000 at the beginning of the third quarter of 2022.
The interesting thing is not so much the short-term availability of these graphics, but the rumors about their performance. According to those data, the hypothetical RTX 4090 will double the performance of the RTX 3090 with the same consumption, something simply amazing that is even hard to believe.
Vertigo Specs
filtration comes from from a well-known Twitter user called “kopite7kimi” who has been right on previous occasions and who this week commented on how effectively these new graphics with “Ada Lovelace” architecture will be presented even sooner than expected: could arrive in mid july.
These are the theoretical specifications of the first three graphics cards of the new RTX 4000 family. The data, beware, is not official.
In addition to that release date forecast, kopite7kimi revealed also several of the key specifications of these graphics and pointed to that performance “2×3090” which would have, for example, the RTX 4090 with Ada Lovelace architecture.
The differences with the current RTX 3000 family are notable. The RTX 4090 will have for example 126 Streaming Multiprocessors (SMs)and although there was talk of 140, the figure is clearly higher than the 84 of the RTX 3090. It will have the same graphics memory (24 GB of GDDR6X) and will consume the same (450 W).
This data is especially interesting because previously it was said that the consumption of this “graphic beast” could rise to 600 W. Expected In addition, this GPU is manufactured with 4nm photolithography and it would also be a PCIe 5.0 graphics: Those two factors could contribute to that promise of performance, and if these data are fulfilled, we will be facing a truly remarkable generational leap.
Image | Reddit
Via | Ars Technica
