Nvidia is now starting to develop next-generation VR graphics cards. It is reported that Nvidia's next-generation VR card will be named GTX 20XX, based on the Pascal 2.0 architecture. This may not be good news for players who have just started Pascal architecture.
Some foreign media broke the news that the new GTX 20XX will also use Pascal architecture, but will upgrade the process to 14nm FinFET. According to reports, the Pascal architecture of the latest process will be used first in the GeForce GTX 2080 Ti, based on an improved version of the GP102. The GTX 2080 Ti will be the first to use 384-bit memory GDDR5X, while the improved GP102 core will integrate 3384 CUDA cores.
The source also said that the second graphics card still uses the GP 102 core, but it will be the GTX Titan Black Box V2 version. The V2 version will integrate a 384-bit bus interface, supporting up to 24GB of GDDR5X memory at a speed of 12 GB/s. V2's 3840 CUDA core frequency is locked at 1600 MHz. Since HBM is not used, its total bandwidth is up to 576 Gb/s.
A few days ago, Oculus just announced the lowered configuration of the VR computer. According to this situation, it seems that the performance of the GTX 2050 and GTX 1050 Ti versions introduced next year will exceed those of the previous GTX 960, and will even be close to that of the GTX 970. Very willing to GTX 2050 and GTX 1050 Ti will become the killer of low-end VR graphics cards, speed up the popularity of VR computers.