Nvidia will take the competition in the graphics card market one step further with the RTX 4000 family. New details have emerged about the flagship model RTX 4090.
Rumors have it that AMD will offer 4x better single-core performance in the RDNA 3 architecture than the previous generation. Wanting to appeal to more customers in the graphics card market, AMD will challenge Nvidia with its flagship model. It is claimed that Nvidia will come up with a similar upgrade to users.
The FP32 value of the Nvidia RTX 4090 model will be 100 TFLOPS
The FP32 performance of the RTX 3090 Ti model will be 40 TFLOPS, and the RTX 3090 model will have a 35.58 TFLOPS value. It is stated that the AD102 GPU, powered by Ada Lovelace architecture, will have FP32 performance exceeding 2 times.
AD102 GPU will have 18,431 CUDA cores to achieve 100 TFLOPS. It is also estimated that the new model will be offered to users with a clock speed of 2.7GHz. It is stated that the Navi 31 GPU developed by AMD will be introduced with a clock speed of 3.0GHz.
AD102 GPU will be powered by cores produced by TSMC with 5nm technology. It is worth remembering that the previous series used cores produced by Samsung with 8nm technology.
Nvidia used 84 Streaming Multiprocessors (SM) in the RTX 3090 Ti. It has been leaked that the company will appear in front of users with 144 SMs on the RTX 4090.
5335
