• Main
  • News
  • GeForce RTX 4080: 12GB model raises some doubts: is it a disguised RTX 4070?

GeForce RTX 4080: 12GB model raises some doubts: is it a disguised RTX 4070?

23.09.2022 113

According to TechRadar's analysis of Nvidia's stated components, the 12GB GeForce RTX 4080, then the base model, would look more like a hypothetical GeForce RTX 4070 in disguise than a card in the same range as the 16GB model.

The reason for this is not so much the lower amount of VRAM, but the chip used, which is not simply a slower version of the same, but something completely different. Incidentally, the 16GB GeForce RTX 4080 uses an AD103 GPU, there whereas the 12GB model uses an AD104 (the GeForce RTX 4090 uses an AD102, for the record).


The AD104 chip has access to DLSS 3 and all the other graphics innovations in the 40 series, but it is significantly inferior to the AD103 in several respects. In particular, there are 2,048 fewer CUDA cores, there is less memory for caching, and the bus is narrower. Techradar has gone so far as to speculate at least 20-25% lower performance than the AD103 chip, which is no small feat considering the starting price of $899.

The performance gap between the GeForce RTX 3080 and 3070 was about 30 percent, which is why many consider the 12GB GeForce RTX 4080 to be a kind of GeForce RTX 4070 in disguise, despite its very high price. Also supporting this view is the name of the chip itself: AD104, considering that the one in the GeForce RTX 3070, based on the Ampere architecture, was called GA 104.

In this case, this would be a decidedly unscrupulous move on Nvidia's part, considering that usually cards with the same name have differences practically only in VRAM, whereas in this case, they appear to be two different products.

What do you think about it? Share your comments.


Popular on BSgolds

Far Cry 6 review

Far Cry 6 review

Far Cry 6 game review on BSgolds. A detailed review of the game Far Cry 6 with a breakdown of the plot,…