First exposure: RTX 40 SUPER series graphics card models and specifications revealed

News on October 30th, the latest news about the upgrade of the RTX 40 SUPER series graphics cards is emerging one after another, and its specifications and parameters are gradually emerging.

Now, kopite7kimi has announced for the first time the GPU core numbers, the number of CUDA cores and the L2 cache capacity of the three models.

RTX 4080 SUPER will use the AD103 full-blood version core, its specific number is AD103-400, equipped with all 10240 CUDA cores, compared to RTX 4080, this number increased by 512. Rumors about video memory have previously mentioned that it will reach 20GB. GDDR6X, however, will cause the memory bit width to be reduced from 256-bit to 160-bit, and the bandwidth will be significantly reduced, so a more reasonable solution seems to be to maintain 16GB GDDR6X.

As for RTX 4070 Ti SUPER, it will have two different core models, namely AD103-275 and AD102-175, equipped with 8448 CUDA cores, which is higher than RTX The 4070 Ti has 768 more cores. In addition, this model will have 48MB of L2 cache, although the memory configuration is not yet clear.

RTX 4070 SUPER uses two different core models, namely AD103-175 and AD104-350. It has 7168 CUDA cores, which is an impressive 1280 cores more than the RTX 4070. In addition, it is equipped with a 48MB L2 cache, but the memory specifications have not yet been announced

It is worth noting that the RTX 4070 Ti SUPER will be the first model to be named both "Ti" and "SUPER" . However, kopite7kimi is skeptical about this naming method, and there may be changes in the future, including adjustments to specifications

The above is the detailed content of First exposure: RTX 40 SUPER series graphics card models and specifications revealed. For more information, please follow other related articles on the PHP Chinese website!