is nvidia Published a blog post Increasing the performance and technical capacity of a new ‘customs’ processor in Nintendo Switch 2. NVIDIA says the new chip-name has not been named, NVIDIA chip performance in the original switch is 10 times higher, though this claim is completely non-specific.
In particular, NVIDIA is not providing any details about how the performance is compared. Moreover, NVIDIA has also said that the new chip AI enables to promote the DLSS operated by the algorithm. Therefore, while going to the NVIDIA marketing of the past, especially the claim that the new RTX 5070 desktop GPU provides comparable performance to the old RTX 4090, comparing the NVIDIA new chip that is running with Old Chip rendering overcome.
Ann Wadeia says new processors benefit from “1,000 engineer years efforts” and features “in each element, and an NVIDIAGPU, which has amazing visual and AI -powered increase, is dedicated to RT core and tensor core.”
However, there is no indication of which NVIDIA GPU is based on a new chip of architecture, nor any details about its CPU core. Ann Vidia has confirmed that the chip has both RT core and tanker core, which at least places it on the surface with the RTX20 series, otherwise known by the touring, in terms of features.
Of course, there are another three NVIDIA architecture that meets this definition, RTX 30 Series or Imper, RTX 40 Series have also named Lovelace, and RTX 50-series is the latest Blackwell Tech.
It is not fully clear from the NVIDIA blog post about what architecture we are dealing here. This is not completely impossible, though it is unlikely that it is a custom GPU generation that does not make a clean map of any PC graphics technology.
However, previous rumors suggest that the chip is based on an Empire GPU that has 1,536 CUDA core, so this is the most likely specification for now. The original switch is not only 256 with the Maxwell period Cada Corps nor the RT or Tensor Core.
Therefore, even if the new console is not completely 10 times the pure rendering performance, it will be much higher, faster in any case. Whatever the reason, one of the most interesting answers from the argument is whether the new chip DLSS 4 will use the latest and much better transformer model for Applicing.
This is much deeper than the previous CNN model, which is used for previous generations to promote DLSS. This new chip is definitely mostly or perhaps less powerful than any desktop GPU that is compatible with DLSS 4 running. So, does the new switch 2 processor transformer model performance to run? We will know … finally!
