NVIDIA has updated its GeForce RTX 40 line of discrete graphics cards and will soon offer SUPER versions of certain models in the series. To announce the GeForce RTX 40 SUPER, Californians will use the most suitable platform for such purposes – CES 2024. The company’s event is already scheduled for January 8. It is worth reminding that NVIDIA is to announce three new models at once – GeForce RTX 4070 SUPER, GeForce RTX 4070 Ti SUPER, and GeForce RTX 4080 SUPER. Although the official specifications of the cards are not announced beforehand, the available leaks allow us to estimate the characteristics of the upcoming novelties.

GeForce RTX 40 SUPER specs

Let’s take a quick look at the changes, starting with the GeForce RTX 4070 SUPER. Like the base model GeForce RTX 4070, this modification will also use the AD104 GPU. However, in this case, 7168 CUDA cores will be activated, compared to 5888 in the base version. That is, the number of computers will increase by 21.7%. A minimal increase in the base operating frequency is also expected. The memory capacity remains the same – 12 GB. This means that the 192-bit bus resolution and an effective 21,000 MHz for DRAM chips will provide a throughput of 504 GB/s.

Despite the fact that the configuration of the memory subsystem will remain unchanged, after a significant increase in GPU compute capacity, we can expect GeForce RTX 4070 SUPER to be close to the performance of GeForce RTX 4070 Ti. At the same time, the power consumption of the SUPER modification will also increase – 220 watts versus 200 watts for RTX 4070. This point will most likely make it impossible to release GeForce RTX 4070 SUPER models with a single 8-pin auxiliary power connector, and the use of 12VHPWR will become mandatory.

GeForce RTX 4070 SUPER is to be presented on January 17. On this day, there will be reviews of the basic modifications of video cards with the NVIDIA recommended retail price (MSRP). The next day, we will have materials on overclocked more expensive adapters.

The GeForce RTX 4070 Ti SUPER is being upgraded to the AD103 GPU. There is also a certain increase in computing power compared to the base model – from 7680 to 8448, i.e. by +10%. Although the relative increase in GPU functional units is no longer so impressive, the main advantage of GeForce RTX 4070 Ti SUPER will be the increase in memory capacity from 12 GB to 16 GB. Accordingly, the bus bit rate will also increase from 192 to 256 bits.

Even if the effective operating frequency of the GDDR6X chips remains unchanged (21,000 MHz), the memory bandwidth will increase by 33% – from 504 GB/s to 672 GB/s. This overall acceleration will obviously close the current performance gap between RTX 4070 Ti and RTX 4080. In addition, 16 GB will allow you to use the 1440p quality settings without restrictions and experiment with 4K gaming more boldly. Therefore, GeForce RTX 4070 Ti SUPER will potentially become the most anticipated update in the series.

The new line of graphics cards will be presented on January 24. On the same day, reviews with practical performance tests of MSRT-compliant models will appear. More expensive tuned versions will be delayed by a day.

GeForce RTX 40 SUPER vs. non-SUPER

The upcoming GeForce RTX 40 SUPER line, the GeForce RTX 4080 SUPER, will be unveiled a week later on January 31. Judging by the available information, this version will receive the least tangible improvements – plus 5% to the number of computing units with a slight increase in operating frequencies. The memory chips will operate at a symbolic effective 23,000 MHz, which will increase the total bandwidth by 2.7% – from 717 GB/s to 736 GB/s. Therefore, we can predict that GeForce RTX 4080 SUPER will improve the performance of the base model by only a few percent. The power consumption level will remain the same (320 W), but we will see what will happen to the recommended price during the official announcement.