Nvidia: RTX 40 SUPER series will be the key core to defining AI PCs

The definition of AI PCs is still yet to be firmly established. Following the releases of related products from Microsoft, Intel, and others, Nvidia, which has gained dominance in AI servers with its AI GPUs, has introduced the GeForce RTX 40 SUPER series GPU. This series includes RTX 4080 SUPER, 4070 Ti SUPER, and 4070 SUPER, which aims to power gaming and become the key core in AI-driven PCs.

Nvidia stated that the GeForce RTX 40 SUPER series, based on the Ada Lovelace architecture, can provide up to 52 shader tera floating-point operations per second (TFLOPS), 121 ray tracing (RT) TFLOPS, and 836 AI TOPS to enhance gaming and creative capabilities. It already supports over 500 RTX games and applications, preparing users for the wave of generative AI applications about to arrive on PCs.

Nvidia TensorRT platform is suitable for high-performance deep learning inference software, including optimization adjusters and an executive environment that reduces latency and improves capacity during inference applications. TensorRT-LLM for Windows is an open-source library that accelerates the inference performance of the latest LLMs. Regarding AI workloads, compared to the RTX 3080 Ti, the GeForce RTX 4080 SUPER generates videos over 1.5 times faster and generates images over 1.7 times faster.

In terms of gaming, the AI-driven DLSS (deep learning super sampling) provides a higher level of immersion. At the same time, generative AI applications like Adobe Photoshop can utilize Tensor cores to enhance productivity and maintain a stable creative workflow. Nvidia Broadcast eliminates background noise and offers a seamless virtual background.

NVIDIA is confident that with RTX SUPER GPUs, users can unlock the full potential of AI on Windows PCs.

Source: Nvidia, compiled by DIGITIMES, January 2024

Article translated by Jack Wu