一份最新爆料梳理了首批四款RTX 40系显卡的参数，包括RTX 4090、RTX 4080、RTX 4070和RTX 4060。
从CUDA规模来看，RTX 4090比RTX 3090多出53%，RTX 4080比3080多出23%、RTX 4070比3070多出25%、RTX 4060比3060多出26%。
此次爆料有两点略显奇怪，一是说RTX 40系显卡改用了台积电6nm，而非5nm。二是RTX 4060的显存规格，居然比3060少了4GB容量，看来换装G6X后，成本、散热压不住，TGP这番增加了30W。
Huang Renxun has predicted that NVIDIA will usher in a large-scale product update wave in the second half of the year, including but not limited to GPU products.
Obviously, this means that the RTX 40 series graphics cards are already on the line.
A new revelation has sorted out the parameters of the first four RTX 40 series graphics cards, including RTX 4090, RTX 4080, RTX 4070 and RTX 4060.
From the perspective of CUDA scale, the RTX 4090 is 53% more than the RTX 3090, the RTX 4080 is 23% more than the 3080, the RTX 4070 is 25% more than the 3070, and the RTX 4060 is 26% more than the 3060.
Based on this conservative estimate, the increase of CUDA and the overall performance of the graphics card should be the same. Of course, in game scenarios such as light pursuit and DLSS, it can be expected that the RTX 40 series will double or even increase by 2 to 3 times.
There are two points that are a little strange in this revelation. One is that the RTX 40 series graphics cards have switched to TSMC 6nm instead of 5nm. The second is the memory specification of RTX 4060, which is 4GB less than that of 3060. It seems that after the replacement of G6X, the cost and heat dissipation cannot be suppressed, and TGP has increased by 30W.
On the same day, another leaker, Kopite7kimi, also gave news that the AD104 core has been tested, and it is expected to be used on the RTX 4070.
There is widespread speculation that NVIDIA will release the RTX 40 series graphics cards as soon as July and go on sale in September.