NVIDIA, as she herself put it, "dissonated" the version of the GeForce RTX 4080 video card with 12 Gbytes of memory. In more conventional terms, the company changed its mind to produce it. The manufacturer also called the release date of the older version of GeForce RTX 4080 with 16 Gbyte of memory on November 16.
NVIDIA presented two versions of GeForce RTX 4080 at the same time as the flagship GeForce RTX 4090 was already on sale at the September 20 event. At that time, many noticed that GeForce RTX 408012GB was different from GeForce RTX 408016GB not only in terms of the amount of operational memory. The older version was built on the AD103-300 graphic processor with 9,728 CUDA kernels, while the younger one received the AD104-400 chip with 7,680 CUDA kernels. Of course, the chips differ in the number of other functional elements.
The difference is very significant, as well as the NVIDIA tests that were recently published, and the younger model was about 24-30% slower than the older one, so many users agreed that it would be much more honest to release the younger model under the name GeForce RTX 4070.
And the NVIDIA decided to listen to the audience's opinion, and the company said that "". It was also pointed out that "" this suggests that the "disclosed" video map actually debuts as GeForce RTX 4070.