Nvidia announced the GeForce RTX 4000 series, which uses the next-generation GPU Ada Lovelace architecture.
The RTX 4000 series uses TSMC 4nm. Ray tracing and DLSS also evolved to the third generation.
The Ada Lovelace installed in this GeForce RTX 4000 series uses TSMC 4nm, and the AD102 GPU, which is the top GPU, is equipped with 76 billion transistors, nearly three times the 28 billion transistors of the conventional GA102 GPU, and CUDA cores.
In this Ada Lovelace, the Tensor Core responsible for deep learning has been evolved to the 4th generation, and the RT Core responsible for ray tracing is updated to the 3rd generation, which will significantly improve the performance of ray tracing and DLSS.
Only a brief explanation about Ada Lovelace was given during the announcement, and it is expected that further information related to that will be released in the future.
In addition, for this GeForce RTX 4000 series, a total of three models, the top model RTX 4090 and the high-end models RTX 4080 16GB and RTX 4080 12GB, are scheduled to appear.
Nvidia GeForce RTX 4090: Performance is double that of RTX 3090
The Nvidia GeForce RTX 4090 has appeared as Nvidia’s top model, equipped with an AD102-300 GPU, 16384 CUDA cores, and a maximum operating clock of 2.52 GHz. which is about 2.3 times the performance that was available on the GeForce RTX 3090. The VRAM is equipped with 24 GB of GDDR6X, the bandwidth is about 1 TB / s, and the performance has significantly improved. Boasts a power consumption of 450W, and the RTX 4090 Founders Edition is equipped with 12VHPWR as a power supply.
The RTX 4090 will be available for $1599 from October 12th, 2022. The RTX 4080 with 12GB will be available for $899, while the 16GB will cost $1199.
