NVIDIA used the pre-CES 2024 event to make a presentation GeForce RTX 4070 SUPER, GeForce RTX 4070 Ti SUPER and GeForce RTX 4080 SUPER, three graphics cards that we already knew quite well thanks to previous rumors and whose specifications have been confirmed. This means that almost all the information we shared with you turned out to be true.

The GeForce RTX 40 SUPER series is presented as intergenerational renewal which comes with a very clear goal, to improve the price-performance ratio of the models, it succeeds, although in the case of the GeForce RTX 4070 SUPER we have something very curious, which is that it will coexist with the GeForce RTX 4070. The other two models Yes, it will replace the existing version, i.e. on the GeForce RTX 4070 Ti and GeForce RTX 4080, which will no longer be available once accumulated stocks are exhausted.

These three graphics cards use Ada Lovelace graphics architecture from NVIDIA, which means that they have all the technological advances that I have already explained to you on many occasions, and that they also offer a high level of efficiency. So that you don’t have to consult other previous articles, I leave you a summary below with the most important keys of the mentioned architecture:

• New programmable shaders with SER technology, which is able to regroup worker threads into blocks or groups that shader engines will be able to dispatch faster and more efficiently. It can increase performance by up to 25% at base load and double the speed of ray tracing operations.
• RT cores of the third generation, which double the speed of calculating ray intersections and triangles and have Displaced Micro-Mesh and Opacity Micro-Mask technologies. The first uses micro-triangular meshes to improve efficiency when working with ray tracing, and the second reduces the load on shaders by more accurately determining the opacity and density of objects.
• The fourth generation of tensor cores, which can work with operations below FP8 and have the same FP8 transformation engine that NVIDIA introduced with the Hopper architecture. Its performance has increased enormously when working with inference, deep learning and artificial intelligence workloads.
• A new generation of optical flow accelerator, which processes the necessary information about the direction and speed of moving pixels in each frame. This allows accurate estimation of the position of each pixel and is the key to achieving a high-quality result with the frame generation that is part of NVIDIA DLSS 3.

At the aesthetic level, three GeForce RTX 40 SUPER they look amazing, thanks to the benefits of the Founders Edition modification, which as our regular readers know is the best thing NVIDIA has done to date in the general consumer graphics sector. In this intergenerational renewal, a touch of black which they gave to the borders and other elements, so that now they have a completely dark.

GeForce RTX 4070 SUPER: Specs, performance, price and release date

• Graphics core AD104 in 5nm.
• 7168 shaders at 1980MHz-2475MHz.
• 224 texturing units.
• 80 raster units.
• 224 tensor kernels of the fourth generation.
• 56 cores of the third generation RT.
• 192-bit memory bus.
• 12 GB GDDR6X at 21 Gbps.
• 36 shader-TFLOP.
• 82 RT-TFLOPs.
• 568 AI-TFLOPs.
• TGP 220W.

GeForce RTX 4070 SUPER has 20% more shaders than GeForce RTX 4070 and according to NVIDIA It is faster than the GeForce RTX 3090. Obviously, it is also much more efficient, as its TGP is only 220 watts, which is 130 watts less compared to GeForce RTX 3090, or 150 watts less if we take the average consumption in games as a reference. He also has an advantage in support frame generationtechnology that significantly improves fluidity in games.

In terms of performance, we can see that the GeForce RTX 4070 SUPER far outperforms the GeForce RTX 2070, and also It is much more powerful than the GeForce RTX 3070 even without resorting to DLSS 3 (frame generation). A very interesting update if we come from one of these two graphics cards, even more so if we keep in mind the difference that creates the generation of frames, and I can confirm that it is very big.

It will be available from January 17 for a price of 599 dollars, the figure in euros should be 649.99 euros including VAT. If only NVIDIA would hurry up and list it at €599.99 can crash the market completely and place him in a privileged position.

GeForce RTX 4070 Ti SUPER: Specs, performance, price and release date

• Graphics core AD103 in 5nm.
• 8,448 shaders at 2,340-2,610 MHz.
• 264 texturing units.
• 112 raster units.
• 264 tensor kernels of the fourth generation.
• 66 cores of the third generation RT.
• 256-bit memory bus.
• 16 GB GDDR6X at 21 Gbps.
• 44 shader-TFLOP.
• 102 RT-TFLOPs.
• 706 AI-TFLOP.
• 285W TGP.

The GeForce RTX 4070 Ti SUPER represents a very important leap over the GeForce RTX 4070 Ti, as it not only increases the number of shaders and other GPU elements, but also jumps from a 192-bit configuration and 12 GB of GDDR6X memory to a 256-bit bus with 16 GB of GDDR6X memory. This is very important because it is generally a much more versatile model and capable of offering excellent performance even in 4K resolution.

To be able to Beat the GeForce RTX 3070 Ti by up to 2.5 times and is more efficient as it has a TGP of 285 watts compared to the 290 watts that a single GeForce RTX 3070 Ti would have. However, the average consumption values ​​in games are much lower and They have about 226 watts, according to data shared by NVIDIA.

The performance graph allows us to see that the difference between GeForce RTX 4070 Ti SUPER and GeForce RTX 2070 SUPER is huge, and it there is a significant jump in raw power between the first and the GeForce RTX 3070 Ti, even if we do not take into account DLSS 3 (frame generation). The generational leap is so great that it is clear It would be worth upgrading from a GeForce RTX 3070 Ti to a GeForce RTX 4070 Ti SUPER.

The launch of the GeForce RTX 4070 Ti SUPER will take place on January 24 and its price will be 799 dollars. In Spain it should be 849.99 euros including VAT. If confirmed, he will have everything to become graphics card with the best price/performance ratio in its line.

GeForce RTX 4080 SUPER: Specs, performance, price and release date

• Graphics core AD103 in 5nm.
• 10,240 shaders at 2,295 MHz-2,550 MHz.
• 320 texturing units.
• 112 raster units.
• 320 tensor cores of the fourth generation.
• 80 cores of the third generation RT.
• 256-bit memory bus.
• 16 GB GDDR6X at 23 Gbps.
• 52 shader-TFLOP.
• 121 RT-TFLOP.
• 836 AI-TFLOPs.
• TGP 320W.

The GeForce RTX 4080 SUPER is the most powerful model of the three presented by NVIDIA. It is designed for 2160p (4K) resolution playback in a completely optimal way even with active beam tracking, it has twice the performance of the GeForce RTX 3080 Ti and it is much more efficient because it has a TGP of 320 watts, i.e. 30 watts less than the GeForce RTX 3080 Ti. If we take the average consumption in games as a reference, the difference is even greater, as the GeForce RTX 4080 SUPER registers values 246 watts.

As we can see in the attached GeForce RTX 4080 SUPER chart sling soup on GeForce RTX 2080 SUPER, and if we compare it to the GeForce RTX 3080 Ti, the difference in favor of the former is still huge. If we add DLSS 3 (frame generation) to the equation, the difference will be so great that It almost looks like we are facing a difference typical of two graphics generations.

The launch of the GeForce RTX 4080 SUPER is scheduled for January 31, and its price will be 999 dollars. At the euro exchange rate and after taxes, that figure is likely to be €1,099 including VAT, meaning it will be cheaper than the GeForce RTX 4080 at launch. If confirmed, it will offer more interesting value than expected in the price-performance ratio.