NVIDIA GeForce RTX 40 series GPUs now official with 4X performance increase, ADA Lovelace Architecture, DLSS 3

During the GTC event, NVIDIA announced the GeForce RTX 40 series GPU. The RTX 4090 is said to be 4x faster than the 3090. Read on for more!

NVIDIA GeForce RTX 4090 and more are now official

NVIDIA GeForce RTX 40 series officially launched with new features and better performance

NVIDIA RTX 40-series boasts 4X more performance

First off, the new Ada Lovelace Architecture debuted on the GeForce RTX 40 series featuring 3rd Gen RT cores. These new RT cores offer up to 2x the throughput of the last gen. This will also speed up the rendering of ray-traced motion blur for fast results and better visual accuracy.

The 4th Gen Tensor Cores accelerate more data types and support Fine-Graned structured Sparsity to deliver more than 2x throughput for tensor matrix operations compared to the last generation. For single-precision floating-point computations, the Ada-based CUDA cores increase processing speed over the GPUs of the previous generation by 200 percent. For graphics processes like 3D model development and compute workflows like desktop simulation for computer-aided engineering, this results in significant performance increases.

The next new feature is the NVIDIA DLSS 3.0. It now combines DLSS Super Resolution, all-new DLSS Frame Generation, and NVIDIA Reflex which is made possible by the abovementioned hardware improvements of the RTX 40 series. This is expected to be 4x faster than straight-up rendering.

According to NVIDIA, the gaming ecosystem has already embraced DLSS 3 at a rapid pace, with over 35 games and applications incorporating the technology, the first of which will go on sale in October. According to the business, DLSS 3 can reduce latency by up to 2X when compared to native.

Other new features include streaming multiprocessors with up to 83 teraflops of shader power, Shader Execution Reordering, Ada Optical Flow Accelerator, Architectural improvements, and Dual NVIDIA Encoders.

NVIDIA GeForce RTX 40-series specs

The NVIDIA GeForce RTX 4090 will feature 24GB GDDR6X VRAM, 384-bit memory interface width, and a 450W TDP. The NVIDIA GeForce RTX 4080 will have a 16GB variant and a 12GB variant as well. With the new tech and features, the RTX 4090 is expected to be 4 times faster than the RTX 3090.

NVIDIA claimed that it constantly produces over 100 frames per second when gaming at 4K resolution. It has 76 billion transistors, 16,384 CUDA cores, and 24GB of high-speed Micron GDDR6X memory. NVIDIA claimed that it constantly produces over 100 frames per second when gaming at 4K resolution. It has 76 billion transistors, 16,384 CUDA cores, and 24GB of high-speed Micron GDDR6X memory.

International price with direct to PHP conversion (no taxes and duties yet):

NVIDIA GeForce RTX 4080 12GB - USD 899 (around PHP 52K)

NVIDIA GeForce RTX 4080 16GB - USD 1,199 (around PHP 69K)

NVIDIA GeForce RTX 4090 24GB - USD 1,599 (around PHP 93K)

The RTX 40-series will start to be available starting October and November. Limited edition Founder's Edition will also be available.

What do you guys think?

Source: NVIDIA