▷ Deep learning super

Deep Learning Super Sampling (DLSS) is one of the most promising technologies in Nvidia's new Turing graphics architecture. This technology builds on the artificial intelligence (AI) capabilities of the company's graphics cards to improve video game performance without increasing raw power. We tell you all about DLSS and how it works.

Index of contents

How does Deep Learning Super Sampling work on the new Turing graphics cards?

Tensor Core are the fundamental element of the Turing architecture for the operation of Deep Learning Super Sampling. Nvidia's Tensor Core are special cores that are designed to speed up the computation of multiple matrices, the math commonly used in deep learning algorithms and other computing scenarios focused on artificial intelligence.

Some of our readers may wonder why Nvidia has decided to bring this enterprise-grade feature to the gaming industry, but the answer is pretty simple. Nvidia has long worked with AI capabilities related to image reconstruction, and has found a way to exploit this in video games.

We recommend reading our post on What is rasterization and what is its difference with Ray Tracing

Nvidia will use DLSS to do high-quality rescaling on games, this means they will render at a lower resolution than the final, resulting in better performance. For example, you could render an image at 2K and then enlarge it to 4K using DLSS capabilities, this results in an image with a quality very similar to a native 4K image, but with much higher performance.

performance

Nvidia's Turing architecture uses its Tensor Core for Deep Learning Super Sampling in games, allowing Nvidia to offer similar levels of image quality as a native resolution display with TAA, while offering a significant performance boost.. This gives DLSS users an increase in performance estimated to be around 35-40%, acting as a kind of "free performance upgrade" for games that support the Deep Learning algorithm.

Nvidia's Tensor Core will be used to increase the clarity of gaming with DLSS, reducing the computing power required to process high-resolution images, offering the industry's first AI- powered performance boost. With Deep Learning, Nvidia will be able to create high resolution images, players will not notice the difference compared to an image rendered at native resolution.

Nvidia has stated that they plan to create other technologies that can use their Tensor cores in video games. When it all comes together, Nvidia's concurrent workflow system will allow more computational work to be completed than ever before, further paralleling the GPU workflow.

With Turing, Nvidia has accumulated more computing power on a single graphics card than ever, while diversifying computing or graphics card infrastructure to enable new features, forging a path in the Deep Learning and Ray Tracing domains in time. real.

Games that will use Deep Learning Super Sampling

The list of video games with support for Deep Learning Super Sampling is still quite small, but it will increase as time goes by. For now the list of compatible games is as follows:

• Ark: Survival EvolvedAtomic HeartDarksiders IIIDauntlessDeliver Us The Moon: FortunaFinal Fantasy XVFractured LandsHellblade: Senua's SacrificeHitman 2Islands of NyneJusticeJX3KINETIKMechwarrior 5: Battle of the WildsSuperheroes: Deadline: Battlegrounds

We recommend reading:

This ends our special article on the new technology Deep Learning Super Sampling, remember that you can share it on social networks so that it can help more users who need it.