The fact that Nvidia will present the GeForce RTX 4000 series based on the new GPU architecture, Nvidia Ada, at the GTC 2022 on September 20 is obvious and is supported both by leaked information and many rumors. Now Nvidia invites you to puzzle with a mysterious tweet.
Nvidia keeps talking about #ProjectBeyond
After a month of back-and-forth speculation on the specifications of the first Nvidia GeForce RTX 4000, the first concrete details on the putative specifications of the expected GeForce RTX 4090 and 4080 have emerged, which will be the first models of the new GeForce portfolio manufactured in TSMC N4 are supposed to provide more substantial figures, although none of them are official yet.
Nvidia has now published a mysterious tweet on Twitter that contains a number code, the meaning of which is already being hotly debated. With the keyboard chosen for the video and the open Steam client in the Windows taskbar, the manufacturer once again refers to the topics of GeForce and gaming.
What is (208) 629-7538 and is it even relevant? A first assumption is that it is about performance, chip size, and the number of transistors. It’s possible, but there’s something else behind it: Anyone who calls the number in the USA will hear the following recorded message: “Tell us, how fast would you like to go?”. The note in the tweet that “the number will self-destruct” also goes with this tape announcement.
At the latest on September 20 at 5 p.m. German time, the date of Nvidia’s CEO Jensen Huang’s keynote, there will be official reports on the new graphics cards.
Two very different GeForce RTX 4080
According to the latest rumors, three high-end graphics cards will appear when the new “Ada generation” is released, but this will lead to a particularly significant gap between the upcoming top model, the “GeForce RTX 4090”, and the two in the model range and from the Nomenclature would lead to the GeForce RTX 4080 placed directly below.
- GeForce RTX 4090/24GB
- AD102-300 with 16,384 CUDA cores
- 24GB GDDR6X @ 384-bit
- 450 – 660 watts TGP
- GeForce RTX 4080/16GB
- AD103-300 with 9,728 CUDA cores
- 16GB GDDR6X @ 256-bit
- 340 – 516 watts TGP
- GeForce RTX 4080/12GB
- AD104-400 with 7,680 CUDA cores
- 12GB GDDR6X @ 192-bit
- 285 – 366 watts TGP
In view of the fact that a “GeForce RTX 4080” offers more than 53 percent fewer computing units than a “GeForce RTX 4090” in the worst case, the questions about possible Ti models do not even arise. The “GeForce RTX 4080 Ti” is obviously planned.
First customs and questionable design
Presumably authentic pictures of the first custom designs of the GeForce RTX 4090 from the board partners are already doing the rounds, but their product packaging still leaves a questionable impression in terms of design.
TechAint 