Yesterday we informed you that the upcoming NVIDIA GeForce RTX 4000 SUPER graphics cards are to debut at the very beginning of January 2024, during the next edition of CES. So far, three card models have been suggested: GeForce RTX 4080 SUPER, GeForce RTX 4070 Ti SUPER and GeForce RTX 4070 SUPER. I admit that until yesterday I did not believe in the existence of the second card mentioned, but new information forces me to verify this opinion.
The logo of the NVIDIA GeForce RTX 4070 Ti SUPER graphics card has been published online. It not only presents a changed appearance of the word SUPER, but also confirms that a card with such an unusual name is actually being created.
Portal X user – @MEGASizeGPU – published a marketing photo on its account, showing not only the new SUPER logo, but also the full name of the GeForce RTX 4070 Ti SUPER graphics card. Similar materials appeared before the premiere of the GeForce RTX 4000 series, and it looks like this time we are also dealing with authentic content. This confirms the existence of a card with a rather unusual name. Thanks to the modified specification, the graphics system will be much closer to the GeForce RTX 4080 model.
|RTX 4090||RTX 4080 Great||RTX 4080||RTX 4070 Ti Super||RTX 4070 Ti||RTX 4070 Great||RTX 4070|
|Core||AD102-300||AD103-400||AD103-300||AD102-175 / AD103-275||AD104-400||AD103-175 / AD104-350||AD104-250|
|VRAM||24GB G6X 384-bit||16GB G6X 256-bit||16GB G6X 256-bit||16GB G6X 256-bit||12GB G6X 192-bit||12GB G6X 192-bit||12GB G6X 192-bit|
|TDP||450 W||320 W||320 W||285 W||285 W||220 W||200 W|
NVIDIA GeForce RTX 4080 Super and RTX 4070 Ti Super are expected to consume the same amount of power as the variants without the Super note
NVIDIA GeForce RTX 4070 Ti SUPER, according to current information, will be equipped with one of two Ada Lovelace cores: AD102-175 or AD103-275. In both cases, we are dealing with truncated cores, offering a total of 8,448 CUDA processors. Compared to the regular GeForce RTX 4070 Ti, the improved version will also offer 16 GB of GDDR6X memory on a 256-bit bus, instead of the previous 12 GB GDDR6X on a 192-bit bus. The TGP coefficient will remain at 285 W, just like in the previous version.
Source: X @MEGASizeGPU