Nvidia took the curtains off of its next-generation RTX 4000 graphics cards last night, with one of them launching as soon as October 12. For now, we are only getting the RTX 4080 and 4090 and their cheaper siblings will be unveiled later this year.
RTX 4090
The RTX 4090 is now Nvidia’s top-of-the-line discrete GPU for gamers and professionals. As the name says, it is a successor to the RTX 3090 and yet it is 2X faster than even the RTX 3090 Ti, according to Nvidia’s claims and up to 4X faster when using DLSS. It is an absolute behemoth of a GPU with 16,384 CUDA cores with boost clocks that go up to 2.52GHz. It has 24 GB of GDDR6X video memory and a 384-bit memory interface.
As expected, you will need a power plant to back up your computer since it has a power draw of 450W. It runs on a single 16-pin PCIe Gen 5 or 3x 8-pin PCIe cables. It will be available for sale starting October 12 at a starting price of $1599.
The RTX 4090 will replace the existing 3090, but the 3060, 3070, and 3080 graphics cards will continue to sell. Lower-end RTX 4000 GPUs are not expected to hit the market before 2023.
RTX 4080
Nvidia has launched two variants for the RTX 4080, one with 16 GB GDDR6X memory and the other with 12 GB. The former has 9728 CUDA cores and a 2.51GHz boost clock with a 256-bit memory interface, while the latter is a slightly cut down version with 7680 CUDA cores, a 2.61GHz boost clock frequency, and a 192-bit memory interface. Nvidia says that the 4080 is 2-4x faster than the RTX 3080 Ti.
The 16 GB model’s power draw is at 320W and the 12 GB variant will take up 285W. The 16 GB RTX 4080 will be available for $1199 and the 12 GB sibling will cost you $899. These two will go for sale in November.
Both of the new graphics cards are running a cut down version of Nvidia’s full size Ada Lovelace GPU built on TSMC’s 4nm process. This GPU brings 3rd generation RT cores, 4th generation Tensor Cores, and support for new features including DLSS 3 and AV1 encoder.
DLSS 3
Deep Learning Super Sampling, better known as DLSS, is Nvidia’s AI-based technology to boost FPS in video games. The company has now announced DLSS 3, which will be exclusive to RTX 4000 GPUs. It uses an Optical Flow Accelerator along with data from sequential frames to create frames on the GPU itself, rather than within the game. This should improve in-game FPS by up to 4 times.
DLSS 3 will also overcome inherent CPU limitations that affect traditional upsampling and image reconstruction methods as the frame rate increases. In layman terms, all kinds of games (including CPU intensive ones) can benefit from this tech, improving frame rates dramatically.
More than 35 games already have support for this feature such as Cyberpunk 2077, Microsoft Flight Simulator, Dying Light 2, and more.
