Faster and more expensive: Nvidia unveils Geforce RTX 4080 and RTX 4090

Nvidia has officially announced the new Geforce RTX 4080 and Geforce RTX 4090. Find all the details here!

Enlarge
Nvidia Geforce RTX 4080 and 4090 Ti will launch on October 12, 2022
© Nvidia
Nvidia boss Jensen Heng officially launched the first Geforce RTX 4000 graphics cards with the new,
GPU based on the Nvidia Ada Lovelace architecture
announced. Nvidia speaks of a “revolutionary performance for gamers and creative people” and a “significant generation change in terms of performance efficiency”. The new series supports DLSS 3 and “stands for a new era of real-time ray tracing and neural rendering that uses AI to generate pixels”.

Enlarge
The Nvidia ADA Lovelace GPU is in the new Geforce RTX 4000 graphics cards
© Nvidia
The new GPUs have 76 billion transistors and are manufactured using the 4-nanometer process. The GPU features a new streaming multiprocessor with 90 shader TFLOPs and twice the performance and energy efficiency compared to its predecessor. Ray tracing uses the third generation of RT cores with 200 RT TFLOPs, which is said to correspond to a 2.8x increase in performance. In “deep learning” 4th generation Tensor cores with 1,400 Tensor TFLOPs (about 5x more than the previous generation with FP8 acceleration).
Nvidia clearly designed its new “Ada Lovelace” architecture to handle ray tracing tasks at top speed. At its core, the GPU uses new “Shader Execution Reordering” to improve rendering of traditional games by 25 percent, with the ability to render ray-traced scenes two to three times faster.
That changes with DLSS 3
In addition to the shader cores, new technologies introduced in Lovelace’s third generation of dedicated ray tracing hardware (like Opacity Micro-Maps) also aim to dramatically accelerate ray tracing performance. But Nvidia’s “RTX” foray doesn’t just rely on the power of ray tracing. Huang also revealed
DLSS 3,
which runs on a new generation of Tensor cores and is set to replace the second-gen version of Nvidia’s AI-enhanced upscaling feature.
While previous DLSS versions used machine learning to generate new pixels and allowed Geforce graphics cards to run games at lower resolutions (for better performance) before adapting to the monitor’s resolution, DLSS 3 generates entirely new frames to accommodate the determine optimal upscaled image quality.
Our tests will then show what that brings. With the RTX 40 series, however, real-time ray tracing could become feasible.
Geforce RTX 4090 available from October 12th and Geforce RTX 4080 from November
the
Geforce RTX 4090
will be available with 24GB of G6X storage for $1,599 on October 12, 2022.

Enlarge
Geforce RTX 4090 will be available from October 12th
© Nvidia
Nvidia promises double and up to four times the performance of the Geforce RTX 4090 compared to the Geforce RTX 3090 Ti. At the unveiling, the Nvidia boss explained that the RTX 4090 offers twice the performance of the RTX 3090 in Microsoft’s Flight Simulator Ti. Tripled performance on Portal RTX and quadrupled performance on Racer X.
“In fully ray-traced games, the RTX 4090 with DLSS 3 is up to 4x faster than the last-gen RTX 3090 Ti with DLSS 2,” said Nvidia. In addition, it is up to twice as fast in current games with the same 450W power consumption. The RTX 4090 has 76 billion transistors and 16,384 Cuda cores. With 24 GB of high-speed Micron GDDR6X memory, the GPU is said to offer enough performance for consistently over 100 frames per second in 4K games.

The Geforce RTX 4080 will be available from November 2022 with 12 GB G6X and 16 GB G6X memory. It starts at $899 and the model with more storage starts at $1,199. Here, Nvidia promises twice to four times the performance compared to the Geforce RTX 3080 Ti.

Enlarge
According to Nvidia, the RTX 4080 should offer this performance
© Nvidia
Specifically, the RTX 4080 has 16GB, 9,728 CUDA cores and 16GB of high-speed GDDR6X memory from Micron. With DLSS 3, it is twice as fast in current games as the GeForce RTX 3080 Ti and more powerful than the GeForce RTX 3090 Ti, with less power consumption. The RTX 4080 12GB, on the other hand, has 7,680 CUDA cores, 12 GB of Micron GDDR6X memory and with DLSS 3 it is faster than the RTX 3090 Ti, the flagship of the previous GPU generation.
The entire family of RTX graphics cards
With the new RTX 4000 models, Nvidia is expanding the family of RTX graphics cards. The entry-level model remains (for now) the RTX 3060, priced starting at $329. The RTX 3070 and RTX 3080 follow. The new top models then include the RTX 4080 with a starting price of $899 and finally the new flagship RTX 4090 with a price of $1,199.
Similar to the RTX 3090 before it, the RTX 4090 is technically a “Geforce” branded graphics card, but it’s not really intended for gamers. Instead, it is aimed more at “prosumers” with a generous 24 GB of GDDR6X memory and the steep price of 1,599 US dollars. So users who use the GPU for work during the day and then want to enjoy high frame rates in games at night.
RTX40 prices are higher than when the RTX 30 was launched
Fears about even higher prices for the RTX 40 series have now been confirmed. The RTX 4090 starts at $1,599, a $99 premium over the RTX 3090. The 16GB Geforce RTX 4080 is $1,199, the same price as the RTX 3080 Ti and $400 more than the RTX 3080 at $700. Finally, the 12GB RTX 4080 starts at $899, a $200 increase over its predecessor.
AMD is already positioning itself with RDNA3
A few hours before Jensen Huang’s keynote address, Radeon CEO Scott Herkelman tweeted that AMD’s next-gen Radeon cards will be RDNA3
to be unveiled on November 3rd,
while AMD Technology Architect Sam Naffziger wrote a blog post touting RDNA3’s huge improvements in efficiency. That’s arguably a direct shot at the presumably high performance numbers that Nvidia’s RTX 40 series will need to run at top speeds.
Brief overview: The new RTX 4000 models in brief
RTX 4090:
– 16384 CUDA arithmetic units
– Boost Clock: 2.52GHz
– 24GB GDDR6X
– Power consumption: 450 W
– MSRP: $1599
– Released: October 12, 2022
RTX 4080 (16GB):
– 9728 CUDA arithmetic units
– Boost Clock: 2.51GHz
– 16GB GDDR6X
– Power consumption: 320 W
– MSRP: $1199
– Release date: November 2022
RTX 4080 (12GB):
– 7680 CUDA calculation units
– Boost Clock: 2.61GHz
– 12GB GDDR6X
– Power consumption: 285 W
– MSRP: $899
– Release date: November 2022
Reading tip: Graphics card comparison 2022 with ranking
Cheaper than ever: high-end graphics cards with RTX 3080 Ti and 3090 (Ti) at the best price