Webmaster's Home (ChinaZ.com) September 21 News: At the 2022 Autumn GTC Conference last night, Nvidia released the GeForce RTX40 series GPU. The RTX40 series adopts the new Ada Lovelace architecture, and the performance of Ada is 4 times that of the previous generation of products, including RTX4090, RTX408016GB, RTX408012GB and other products.
According to the official data released by Nvidia, the RTX40 series GPU has a series of new technological innovations, including:
● Streaming multiprocessor has shader capabilities up to 83TFLOPS, and its throughput exceeds that of the previous generation of products.
● The effective ray tracing computing power of the third-generation RT Cores reaches 191TFLOPS, 2.8 times that of the previous generation of products.
● The fourth generation Tensor Cores has FP8 tensor processing performance up to 1.32 Petaflops, which is 5 times faster than the previous generation's acceleration performance using FP8.
● shader execution reorder (SER) improves execution efficiency by rescheduling the shader load on the fly, thereby better utilizing GPU resources. As a major innovation as the out-of-order execution of CPU, SER brings up to 3 times performance improvements to ray tracing, with an overall gaming performance improvement of up to 25%.
● Ada optical flow accelerator brings a 2x performance improvement, allowing DLSS3 to predict motion in the scene, allowing neural network to improve frame rate while maintaining image quality.
● Architectural improvements are closely combined with TSMC4N custom process technology, achieving a performance-power consumption ratio of up to 2 times.
● Dual NVIDIA encoder (NVENC) reduces output time by up to half and supports AV1.
Among them, the RTX4090 uses DLSS3, has 76 billion transistors , 16384 CUDA cores and 24GB high-speed Micron GDDR6X video memory, and continues to run at more than 100FPS in 4K resolution games. Compared with the previous generation of RTX3090Ti using DLSS2, the performance improvement can reach 4 times. In modern gaming, the RTX4090 has up to 2x performance while maintaining the same 450W power consumption.
RTX4090 will be available on October 12, with a price starting from 12,999 yuan.
In addition, NVIDIA also announced the launch of RTX4080 with two configuration versions.
RTX408012GB has 7680 CUDA cores and 12GB Micron GDDR6X video memory.
RTX408016GB has 9728 CUDA cores and 16GB high-speed Micron GDDR6X video memory, and its performance in modern games can reach twice that of the GeForce RTX3080Ti. The two configuration versions of
RTX4080 will be launched in November, with prices starting from 9,499 yuan and 7,199 yuan respectively.
