There is a lot of talk about AI in laptops in the context of telling us that the NPU marked a very important “revolution”, but the truth is that this revolution already happened a few years ago with the arrival of the first GPUs. from equipped consumption with specialized AI hardwareGeForce RTX 20 from NVIDIA.

Don’t get me wrong, the NPU is an interesting solution because it is specialized AI hardware that has very low power consumption, but then its real capabilities very limitedand these They were not real ancestors of this revolution called AI in laptops, it was actually the GPU, and as I said, NVIDIA’s Turing was the cutting edge.

NVIDIA’s revolution in the world of artificial intelligence applied to consumer PCs and laptops has been so spectacular that today, six years after the arrival of GeForce RTX 20, some of its most important AI technologies are being introduced to consoles and being adopted by users as if they were something incredible.

The best example is the recently announced PS5 Pro, which uses PSSR, a resizing technology that uses hardware-accelerated AI to improve image quality and performance. Sounds good, right? This technology has been available in notebooks with NVIDIA RTX GPUs for six years, and since then it has improved so much that today, the best intelligent scaling technology out thereboth for performance scaling and image quality.

AI in laptops: the importance of GPUs

Let’s stop at the numbers for a moment. The NPU can achieve maximum performance 50 TOPswhich is a number significantly higher than previous generation solutions, which could barely reach 16 TOP. However, this level of performance is still very low when compared to, for example 321 TOPs offered by the GeForce RTX 4070 used by the ASUS ROG Zephyrus G16.

The difference in performance is so great that the GeForce RTX 4070 It delivers nearly 6.5 times the performance of the most powerful NPU which currently exists on laptops. This simple comparison will help you understand why I say that the magic when talking about AI in laptops is in the GPU and not the NPU.

A notebook equipped with a GeForce RTX 40 can train AI models up to 30 times faster multiply the frame rate in games by five and is able to generate images up to 13 times faster. These data do not refer to the most powerful model, but were obtained from the modest GeForce RTX 4060, a mid-range model that can be found in notebooks under 1,000 euros, which means that the result with the GeForce RTX 4070 will be equal. better.

It is clear that power matters, but it is not everything. It would be useless for us to have 321 TOPs if we don’t count it with an application ecosystem and the necessary support so we can take advantage of it, and that’s where NVIDIA has done its homework brilliantly over these six years, knowing how to balance the gaming side and the more professional side of AI in laptops.

AI in games: higher performance and advanced technologies

NVIDIA has been a great pioneer in the use of AI applied to games. A major breakthrough occurred in 2018, when DLSS Super Resolution technology was born, a smart scaling which utilizes hardware acceleration via GeForce RTX GPU tensor cores.

By combining specialized hardware with AI algorithms, NVIDIA has been able to create intelligent scaling technology that can achieve triple gaming performanceand the best part is that it does so while maintaining superior image sharpness and stability than other types of upscaling.

NVIDIA’s next step in the world of artificial intelligence applied to gaming was DLSS 3 Frame Generation. It was developed as a complement to DLSS Super Resolution and its concept was very successful because it is a technology that generates a completely independent additional frame on the GPU for every two frames rendered traditionally.

This technology uses to generate each of these frames AI and hardware acceleration via Optical Flow Accelerator. This allows you to calculate the movement and direction of the pixels, which is essential to ensure that the position of the pixels in each generated image is correct, that it matches the natural movement of the image and does not introduce errors or graphical problems. .

The value of artificial intelligence applied to gaming in these technologies is so great that the combination of NVIDIA DLSS Super Resolution and NVIDIA Frame Generation we can improve the frame rate per second up to five times, and without having to give up a good level of sharpness and graphic quality.

But that’s not all, NVIDIA continues to find new ways to innovate in the world of AI applied to games, and its latest major advancement in this regard has been DLSS Ray Reconstruction. This technology is defined as a Accelerated intelligent noise limiter on tensor kernelswhich is able to distinguish between good and bad pixels, stores more information and can distinguish different ray tracing effects.

Not only with DLSS Ray Reconstruction Improves image quality by using full ray tracingbut in addition, performance is also improved by directing this technology to traction cores.

AI in professional applications and tools: study, create and work at a different level

The GeForce RTX 40 GPU can do its magic with gaming technologies, but did you know it also has something to say in professional applications and tools? Well, that’s right, all those technologies that we just saw were also integrated into this type of application, and this makes an important difference.

For example, NVIDIA Omniverse supports NVIDIA DLSS Super Resolution and DLSS Frame Generation, Autodesk Arnold supports hardware-accelerated ray tracing and DLSS 3, and D5 Render offers full support of the NVIDIA ecosystem, including hardware-accelerated ray tracing and the entire set of technologies that make up DLSS 3.5 (DLSS Super Resolution, DLSS Frame Generation and DLSS Ray Reconstruction).

Leading rendering, modeling, and content creation applications are ready to harness the potential of AI through hardware-accelerated NVIDIA GPUs, as well as the most well-known and widely used video editing solutions such as DaVinci Resolve and Adobe Premiere Pro, among others.

The value of the GeForce RTX 40 GPU does not end there. We can enjoy ourselves with him Stable diffusion with TensorRT acceleration from NVIDIA, which will allow us to generate images much faster, and we’ll be able to unleash our most creative side by using the AI ​​in NVIDIA Canvas to transform simple virtual brushstrokes into highly realistic landscapes.

NVIDIA hasn’t forgotten streamers either. If you want to better reach your audience, you can use the value of AI and the potential of your GeForce RTX 40 GPU through NVIDIA Broadcast, a completely free application to create your own home studio and enjoy advanced features through hardware-accelerated artificial intelligence, such as background noise-free and echo-free audio and higher image quality without affecting performance.

You can also train your own chatbots, you will have fun your personalized LLM locally thanks to ChatRTX, You’ll have everything you need to develop your own AI applications with the RTX AI Toolkit, which includes a comprehensive SDK, and you’ll have access to hundreds of pre-trained AI models and safely modify them with your own data. With the GeForce RTX 40, you set the limit.