NVIDIA just announced its new flagship graphics cards, the RTX 40 series.
The new NVIDIA GeForce RTX 40 series are powered by the brand’s new TSMC 4nm Ada Lovelace architecture. They are rigged with 4th generation Tensor cores, 3rd generation RT cores, and support for the latest DLSS 3 and AV1 encoder.
Compared to its predecessor, the RTX 4090 has 83 TFLOPS. It also has 2x better shader performance and ups game frame rates by 25%, all thanks to the new Shader Execution Reordering.
Meanwhile, the RT performance on the top-of-the-line GPU is around 191 TFLOPS, versus the 78 TFLOPS of the RTX 3090 Ti.
The RTX 40-series chips also have two 8th generation NVIDIA Encoders (NVENC). This new feature can both decode and encode AV1. The ShadowPlay capture resolution has also been improved to 8K at 60fps with HDR.
One of the chips’ biggest features, the DLSS 3, use Optical Flow Accelerator and sequential frames to improve game frame rates up to 4x. Upon launch, the game will support over 35 games.
Now, the specs differences. First, we have the NVIDIA RTX 4090, which has 16384 CUDA cores with a boost clock speed of up to 2.52GHz. It then has 24GB GDDR6X memory, 384-bit memory, 450W power, and can be powered by either one 16-bit PCIe Gen 5 or three 8-pin PCIe cables.
Next stop, we have the NVIDIA RTX 4080 GPU, which has 9728 CUDA cores, a 2.51GHz boost clock, 320W power, 16GB GDDR6X memory, and a 256-bit interface.
It has a second variant, the RTX 4080 12GB version, which has a 192-bit interface, 285W power, 7680 CUDA cores, and a 2.61GHz boost clock.
The NVIDIA RTX 4080 12GB has a price of USD899 (around Php52,000) while the 16GB is at USD1,199 (around Php69,400) and will be available this November. Meanwhile, the RTX 4090 is at USD1,599 (around Php92,500) and will be sold in select markets starting October 12.