Nvidia has accelerated its work for the RTX 50 series, which is expected to be introduced next year. Some technical features of the Nvidia RTX 50 family, which is stated to be released with the Blackwell architecture, have emerged.
Features of the Nvidia RTX 50 family become clear
According to the leaked information, Nvidia will include transistors produced by TSMC with 3nm technology in its new generation graphics cards. Accelerating 3nm production in the fourth quarter of 2022, TSMC announced that production costs increased by 25 percent compared to 5nm technology. It’s worth noting that the Nvidia CEO visited Taiwan to talk to TSMC about securing early 3nm technology for next-gen graphics cards.
The company will appear with a monolith structure in the RTX 50 series. Renowned analyst Kopite7kimi claimed that the Blackwell architecture is designed for both the end user and the server.
According to the report published by RedGamingTech, Nvidia will come up with significant changes in the CUDA architecture for Blackwell graphics cards. Allegedly, the company will optimize the hardware used for visual enhancement and RayTracing.
Samsung is known to be working on brand new GDDR6W DRAM, which should provide even better performance and higher capacities. However, Nvidia’s use of GDDR7 memory technology is another important leak about the RTX 50 family. It is worth noting that data transfer rates up to 36 Gbps can be achieved with GDDR7 technology.
It is among the rumors that the Blackwell GeForce RTX 50 series uses PCIe Gen 5 interface and offers a clock speed higher than 3GHz. Finally, new generation graphics cards are said to be 2 times better than Ada GPUs in terms of performance.