Nvidia has officially announced its newest graphics card (GPU), the RTX 40 series. This GPU comes in two models, consisting of the RTX 4090 and RTX 4080.
These two graphics cards are designed with the third generation “Ada Lovelace” architecture, which is the term Nvidia uses in designing GPUs capable of supporting gaming performance, professional graphics, AI and powerful computing performance. The new RTX 4090 goes on sale on October 12, while the RTX 4080 goes on sale from November.
The company also says the RTX 40 series graphics cards come with new Nvidia ShadowPlay support which is claimed to be capable of running gameplay up to 8K resolution at 60 fps and HDR. Nvidia is also using the latest Encoder (NVENC) with AV1 encoding support and increased efficiency for live streaming.
Speaking of specifications, the RTX 4090 uses 24GB of GDDR6X memory which Nvidia claims is 2-4 times faster than the RTX 3090 Ti, but with the same power consumption. They recommend using a power supply of at least 850 watts for systems with Ryzen 5900X processors.
Inside the RTX 4090 are stored 16,384 CUDA Cores, a base clock of 2.23GHz that can be launched up to 2.52GHz, 1,321 Tensor TFLOP, 181 RT-TFLOP, and 83 Shader-TFLOP. 2022).
NVIDIA RTX 4080 Series Specifications
Meanwhile for the RTX 4080, there are two variants offered by Nvidia. The first is the variant with 12GB of GDDR6X memory and the second is the variant with 16GB of GDDR6X memory. Both are claimed to be 2-4 times faster than the RTX 3080 Ti.
In addition to different memory support, the technical specifications of the two RTX 4080 variants are also different. The 12GB Nvidia RTX 4080 GDDR6X is equipped with 7,680 CUDA Cores, clocked at 2.31 GHz to 2.61 GHz, 639 Tensor-TFLOPs, 92 RT-TFLOPs, and 40 Shader-TFLOPs. This graphics card runs with a minimum power supply of 700 watts.
The 16GB RTX 4080 GDDR6X is equipped with 9,728 CUDA Cores, clock speed of 2.21 GHz to 2.51 GHz, 780 Tensor-TFLOPs, 113 RT-TFLOPs, and 49 Shader-TFLOPs. Nvidia recommends a minimum power supply of 700W for the 12GB variant and 750W for the 16GB variant.
“For the Founders Edition RTX 40 graphics card, we continued to improve the Dual Axial Flow Through system, increasing fan size and blade volume by 10% and adding a 23-phase power supply,” explains Andrew Burnes, Nvidia technical marketer.
NVIDIA Introduces DLSS 3
Powered by a 4th generation tensor core and a component called Optical Flow Accelerator, DLSS 3 introduces a new technique, called Optical Multi-Frame Generation. Optical Multi Frame Generation will enable DLSS 3 to generate information in the form of new frames, and not just pixels.
This is claimed to provide a large framerate increase, both in GPU-bound and CPU-bound conditions. NVIDIA also mentioned in their press release that there will be many popular games that will get DLSS 3 support, it is noted that currently there are around 35 game titles planned to get the DLSS 3 update, such as Cyberpunk 2077.
Regarding the price, the RTX 4090 is available on October 12 at a price of USD $1,599. While the RTX 4080 12GB costs USD $899 and the 16GB version costs USD 1,199 (Rp 18 million).
Both RTX 4090 and RTX 4080 will be available from various AIB(Add-in Board) partners such as ASUS, Colorful, Gainward, Galax, Gigabyte, Inno3D, MSI, Palit, PNY and Zotac, and will be available in stock clock configurations as well as Custom factory -overclocked cards.