50% faster architecture for classic 3D rendering, twice the performance in raytracing and three times more powerful for calculations relating to artificial intelligence compared to the Ada Lovelace cards of the previous generation: this is in summary what Blackwell promises us. NVIDIA has clearly spared no expense in designing these new GeForce RTX 50 Series, cards that according to its CEO Jensen Huang represent a revolutionary step in computer graphics, with exceptional performance and efficiency for modern graphics and AI applications.
GeForce RTX 50 Series: artificial intelligence in every way
In its most complete version, the architecture Blackwell consists of 92 billion transistors which perfectly illustrate the enormous computing power available with this generation of graphics cards. In addition to physical simulations and 3D rendering, these capabilities will enable efficient and rapid processing of complex tasks in the areas of neural networks, linguistic modeling and real-time data analysis.
The use of GDDR7 memory also makes it possible to double the available bandwidth compared to the GeForce RTX 40 Series and therefore reduce loading times and improve responsiveness, essential conditions for efficiently processing the large quantities of data required for the rendered in 8K, virtual reality or AI.
Also read: Rubin: NVIDIA’s next generation of AI chips already on track
Artificial intelligence is also at the heart of these GeForce RTX 50 Series : the new generation of Tensor Cores can directly process neural networks, offering new possibilities in terms of neural compression of textures or neural shading of materials. This reduces memory requirements while improving the simulation of material properties and pixel generation. The massive use of AI also helps optimize efficiency by avoiding unnecessary calculations, accelerating the generation of complex 3D scenes.
NVIDIA does not forget its technologies dedicated to gamers by introducing the DLSS 4 and multi-frame generation. In practice, these new GeForce RTX 50 Series – and only them – are capable of generating up to three additional images thanks to artificial intelligence, which makes it possible to obtain a framerate up to eight times higher.
Working in unison with NVIDIA Reflex to maintain low control latency, DLSS 4 will be available on nearly 75 games as soon as the first RTX 50 Series are released. A more advanced version of Reflex will soon see the light of day; it will introduce a new functionality called Frame Warp promising a 75% reduction in latency.
Four models at the moment, entry and mid-range will wait
As usual, NVIDIA has chosen to launch the most powerful (and most expensive) models first. At the top of the range, the RTX 5090 is designed for gamers and professionals who need maximum performance and functionality. It is priced at $2000, but in return offers power never before seen on the general public market.
At 1000 dollars, as much as an RTX 4080 Super, the GeForce RTX 5080 displays a better performance/price ratio than its big sister; it should satisfy gamers and easily run the most recent titles in 4K with maximum graphics settings. These two graphics cards will be the first to be sold, with an expected arrival on shelves by the end of the month.
The GeForce RTX 5070 Ti and the RTX 5070, marketed at $749 and $549 respectively, will not arrive until next February. Note in passing that the RTX 5070 Ti is the only one of these four models which will not be offered in the Founders Edition version.
???? To not miss any news from 01net, follow us on Google News and WhatsApp.
