Following the release of the next-generation GeForce RTX 4090 GPU, Nvidia CEO Jen-Hsun Huang confirmed to reporters that the company is removing its NVLink bridge design and that removing the NVLink connector frees up more space to handle more AI operations.
“The reason we took [NVLink] off is that we need I/O for other things, so we’re using that area to cram in as many AI processors as possible,” Jen-Hsun Huang explained of the reason for axing NVLink.
For gamers, creators and workstation users who want to crossfire multiple cards with the latest RTX 4000 series, NVIDIA’s removal of NVLink also enables similar support with PCIe Gen 5.
“Because Ada [Lovelace] is based on PCIe Gen 5, we now have the ability to achieve fast enough point-to-point transfers that it’s a good deal,” added Jen-Hsun Huang.
Compared to PCIe 4, PCIe 5 doubles the bandwidth with direct access to CPU lanes, providing exceptional speed benefits for storage, GPU and memory.
In addition, the RTX 4090 and RTX 4080 graphics cards only have HDMI 2.1 and DP 1.4a interfaces, not the latest DP 2.0 interface. NVIDIA has explained this in a community Q&A, stating that the DP 1.4 interface is now capable of supporting 8K 60Hz output.
Given the high price of the RTX 4090 at launch, Jen-Hsun Huang also declared that “Moore’s Law is dead, the foundry price of 12-inch wafers today is so much more expensive than before that the idea of falling chip prices is a thing of the past.
He said the entire supply chain will increase in price, and the huge inter-generational performance gains make it impossible to meet the “same performance for half the money” condition, or twice the performance for the same money, as in the case of the last generation RTX 3000 series.
“NVIDIA’s $899 GPU or $1599 GPU, compared to the performance a year ago, two years ago, at the same price point, we have much better performance with Ada Lovelace.”
The NVIDIA RTX 4090 flagship graphics card, equipped with 16384 CUDA + 24GB of memory, with a suggested price starting at $12,999, available on October 12.
According to the introduction, RTX 4090 is the world’s fastest gaming GPU. performance, Nvidia said in full ray tracing games, compared with the previous generation of RTX 3090 Ti with DLSS 2, RTX 4090 with DLSS 3 performance improvement of up to 4 times. In modern games, the RTX 4090 delivers up to 2x the performance increase while maintaining the same 450W power consumption.
With 76 billion transistors, 16,384 CUDA cores and 24 GB of high-speed Micron GDDR6X memory, the RTX 4090 consistently runs at over 100 FPS in 4K resolution games.
The GeForce RTX 4090 will be available in China in both standard and overclocked editions from top graphics card vendors such as ASUS, Seven Rainbow, Kongsheng, Shadowcharger, Gigabyte, Image, MSI, and Sotec. OEM products will also be available from Acer, Alien, Asus, Dell, HP, Lenovo and MSI. In addition, NVIDIA will also launch a limited FE version of the RTX 4090.
