[ad_1]
After several months of speculation and rumors, Nvidia officially unveiled the first video cards based on the Ampere graphics architecture: GeForce RTX 3090, RTX 3080 and RTX 3070. Company founder and CEO Jen-Hsun Huang casually touched on a number of new technologies and “chips” of GeForce software, but the greatest attention is all given to the next generation of 3D accelerators.
Nvidia’s new gaming flagship is the GeForce RTX 3090, which replaces the GeForce RTX 2080 Ti. According to Santa Clara, the Ampere architecture has a 1.9x improvement in performance per watt over Turing, as well as doubled ray tracing.
Nvidia Ampere GPUs are produced by South Korean giant Samsung using 8nm technology. Note that previously for flagship GPU production, “green” preferred TSMC factories. The graphics cores also feature PCI Express 4.0 and HDMI 2.1 support.
Nvidia itself refers to the following as the main features of Ampere 3D accelerators:
- New streaming multiprocessors: These are the building blocks of the most powerful and energy-efficient GPU, offering twice the speed of FP32 operations compared to the previous generation and 30 Shader-TFLOPS processing power.
- Second generation RT cores: New dedicated RT cores are twice as powerful as the previous generation, enabling simultaneous ray tracing with shading and computation with 58 RT-TFLOPS throughput.
- Third generation tension cores: New dedicated tensor cores, up to twice as powerful as the previous generation, accelerate and optimize AI-based technologies like Nvidia DLSS with 238 Tensor-TFLOPS processing power.
- Nvidia RTX IO: Fast loading and unzipping of game assets to the GPU, I / O up to 100 times faster than traditional storage and HDD APIs. Combined with the new Microsoft DirectStorage API for Windows, RTX IO offloads dozens of CPU cores to an RTX GPU for faster frame rates and near-instant game loading.
- World’s fastest graphics memory: Nvidia Partners with Micron to Create the World’s Fastest Discrete Graphics Memory for the RTX 30 Series: GDDR6X. The data transfer rate reaches 1TB / s, which improves performance in games and applications.
- Next-generation process technology: Samsung’s new custom Nvidia 8N process technology offers higher transistor density and better energy efficiency.
At the heart of the GeForce RTX 3090 is the GA102 crystal with 28 billion transistors. This graphics card is equipped with 24 GB of GDDR6X memory and according to Nvidia, it is suitable for games with 8K resolution and Titan RTX “up to 50% faster”. It will go on sale on September 24 with an MSRP of around 50.2 thousand hryvnia (or $ 1,499 for the US market).
The GeForce RTX 3080 is a notch lower and, according to Nvidia, offers twice the performance of the GeForce RTX 2080. It has 10GB of GDDR6X memory in its arsenal and the GeForce RTX 3080 will be available for purchase starting September 17. . In stores abroad, the device will be available at a price of $ 699, and in Ukraine – 23.5 thousand hryvnia.
Finally, the GeForce RTX 3070 should replace the GeForce RTX 2080 Ti at half the price and avoid the GeForce RTX 2070 on average by 60%. The video adapter is packed with 8GB of GDDR6 memory and, according to the chip manufacturer, is “an excellent choice for 4K and 1440p gaming.” The new product will be available in October at the MSRP of 16.8 thousand UAH ( $ 499 for the US).
Detailed specifications for the first GeForce RTX 3000 series graphics cards will be announced shortly.
