Before the launch of next-Gen graphics cards from the Nvidia GeForce RTX 4000 series lineup based on Ada Lovelace architecture, it seems like rumors and reports aren’t going to stop.
Today, we have a new one that refers to two variants of the GeForce RTX 4080 with different VRAM memory. The GeForce RTX 4080 16 GB and GeForce RTX 4080 12 GB could arrive simultaneously, as confirmed by two trusted sources at this point.
Two cards, same name, different performance, a different memory. That seems to be the mantra that NVIDIA will follow and with which it intends to fill some gaps in the market.
At the last minute, Nvidia saw something that we still do not see, and instead of launching a GeForce RTX 4070 Ti, the gap will be filled by the 12 GB version, but why?
4080 12G GD6X 10layer PCB aic reference design
4080 16G GD6X 12layer PCB aic
4090 24G GD6X 14layer PCB aic reference design
note：4080 12G has a completely different pcb design from 4080 16G
— MEGAsizeGPU (@Zed__Wang) September 5, 2022
The leak comes to us from MEGAsizeGPU, which points to PCBs pretty fine. The GeForce RTX 4080 12 GB variant has a PCB of lower build quality with 10 layers to its credit, being a reference model in terms of its design.
The GeForce RTX 4080 16 GB will have a 12-layer PCB with the exact blueprint and design target as the GeForce RTX 4090, except that the latter increases its layer count to 14. The problem is that the GeForce RTX 4080 16 GB has a different PCB than the 12 GB version, implying an R&D expense from the manufacturers to bring it to market.
We already know some specs, but the corresponding relationship between specs and models has not been determined.
There is a possibility,
PG141-SKU340/341=RTX 4080 12G;
— kopite7kimi (@kopite7kimi) September 5, 2022
We do not know how many Shaders NVIDIA will cut and if it is simply possible to do so because from the outset, they lose two 32-bit VRAM controllers and a good bandwidth that would not be compensated if this GeForce RTX 4080 12 GB kept the 9,728 cores same as the RTX 4080 16GB.
As there is no chip diagram, we cannot speculate too much but give a perspective; it may go from 76 SM to 70, which would give a result of 8,960 Shaders, almost reduced by 5-8%, while the lost bandwidth would correspond to -25% if what we have discussed is confirmed.
Given the alleged data, it was logical to think that NVIDIA would make a move with the alleged GeForce RTX 4070 Ti later. Still, it seems that AMD will hit hard and right in the middle since there is a more than significant “gap” in performance between GeForce RTX 4080 and GeForce RTX 4070 (9,728 vs. 7,680).
The Radeon RX 7800 XT may be a tougher GPU to beat than NVIDIA expected, and the backlash hasn’t come long. At the moment, it is pure speculation, perhaps this last-minute movement also has a lot to do with the delay suffered, in principle, for the official presentation, or this card may arrive a little later, who knows, place your bets because this is getting hotter and hotter.
Thank you! Please share your positive feedback. 🔋
How could we improve this post? Please Help us. 😔