{"id":584,"date":"2022-11-12T11:01:55","date_gmt":"2022-11-12T06:01:55","guid":{"rendered":"https:\/\/blog.smarteduverse.com\/?p=584"},"modified":"2022-11-12T11:01:55","modified_gmt":"2022-11-12T06:01:55","slug":"what-is-how-does-nvidia-dlss-work-deep-learning-super-sampling","status":"publish","type":"post","link":"https:\/\/codeverse.uk\/index.php\/2022\/11\/12\/what-is-how-does-nvidia-dlss-work-deep-learning-super-sampling\/","title":{"rendered":"What is and How NVIDIA DLSS Works?"},"content":{"rendered":"\n

We have witnessed significant advancements in gaming graphics fidelity over the last few years. The strain on hardware is greater than it has ever been, with more intricate textures, reflections, and sceneries than ever before. This rise in game-related hardware requirements has sparked a number of performance-enhancing improvements. Among these significant developments is Nvidia DLSS<\/a>.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Early in 2019, with the introduction of the Nvidia RTX 2000<\/a> series, the first RTX lineup of GPUs, DLSS was initially introduced. The goal is to double down on the hardware-based speed improvement by using the greater AI capabilities of the RTX GPUs to supercharge frame rate through software optimizations. Support was initially only available for the most recent Battlefield game, but happily, in the three years afterward, the list of compatible games has expanded quickly. The majority of the newest games available today, such as Fortnite, COD Vanguard, Forza Horizon 5, and Battlefield 2042, support DLSS.<\/p>\n\n\n\n

DLSS: What Does It Mean?<\/h2>\n\n\n\n

Deep Learning Super Sampling is referred to as DLSS. The technology makes use of deep learning AI<\/a> to upscale games from their native resolution, which is lower, to the resolution of your display. In essence, this provides you the visual quality of the greater display resolution while maintaining frame rates that you would see at a lesser resolution. The highest visual quality is then achieved and the process\u2019s noticeability is minimized by using AI-accelerated super sampling, a sort of anti-aliasing used to smooth jagged edges in video games.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Now, the free frame rate might seem too wonderful to be true. What\u2019s the catch? The quality of DLSS 1.0 was the catch; there were unmistakably ragged edges and a softer picture. But Nvidia has actually succeeded with DLSS 2.0, developing a revolutionary piece of gaming technology. The findings speak for themselves and help AMD fall behind Nvidia despite having GPUs with the same amount of raw power. It\u2019s all about optimizing the software and the games.<\/p>\n\n\n\n

The load generated by conventional anti-aliasing techniques can be fairly considerable. A picture is rendered out higher than native using a technique called Super Sample Anti-Aliasing<\/a>, or SSAA, and then scaled down. The problem with implementing it this way is that performance will suffer significantly. To produce finer colors and depth, each pixel is basically scaled up by two or four times its original resolution.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Is DLSS Nvidia exclusive?<\/h2>\n\n\n\n

Yes. In an effort to compete with DLSS, AMD announced their Fidelity FX Super Resolution<\/a> technology, however, there is still work to be done. The major factor for Nvidia\u2019s DLSS performance gain is the company\u2019s \u201cdeep learning\u201d secret ingredient. Nvidia has used machine learning to teach AI models to recognize and analyze information from high-resolution scans. This will enable your game\u2019s anti-aliasing to fill in the data using the AI model. This offers excellent super-sampling without the associated overhead. Additionally, DLSS performs this offline, reducing performance loss and boosting frame production.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Alternative Solutions<\/h2>\n\n\n\n

DLSS now offers amazing metrics and gives lower-end GPUs access to greater resolution. There are DLSS substitutes available, including the previously mentioned AMD Fidelity FX<\/a>. When compared to Nvidia\u2019s offering, Super Resolution offers stiff competition. Games often function at a similar level, and Nvidia\u2019s 10 series GPUs have this option as well. As a future feature for Intel Arc Alchemist GPU<\/a>s, XeSS<\/a>, an alternate method of super sampling, is also being developed by Intel. It\u2019s encouraging to see manufacturers offering software ways to boost performance given how often new graphics cards are launched.<\/p>\n\n\n\n

What Games and Cards Can It Be Used In?<\/h2>\n\n\n\n

Although DLSS is fantastic for higher resolutions, there is a catch. DLSS is not accessible for those of you still running GTX 10 series graphics cards. The RTX 2000 and 3000 series GPUs\u2019 Tensor cores are used by DLSS, making them an absolute must for using the technology. Though the new Tensor cores offer higher per-core performance, DLSS will run better on the 30 series GPUs, which still outperform the RTX 2000 series. Who, after all, would refuse a frame rate that is free? Even while the Steam Hardware Survey indicates that GTX 1060 cards are the most popular on the market, it is regrettable that DLSS is not supported on earlier-generation cards. However, the 10-series\u2019 absence of Tensor Cores makes this situation even worse.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

DLSS is extensively embraced in video game titles. The list has only expanded since Nvidia released DLSS, and the majority of well-known games now support the technology, although there are still some gaps. You may look at a table of the titles that Nvidia supports to get a sense of what else is available. Don\u2019t give up if your favorite game isn\u2019t available; Nvidia may support it in the future.<\/p>\n\n\n\n

Measures of performance<\/h2>\n\n\n\n

Three DLSS-compatible titles are listed below, along with the frame rates you may anticipate from each game when utilizing the midrange card, the 3060Ti. To truly demonstrate the potential of DLSS, the resolution and graphics settings are adjusted to 1440p.<\/p>\n\n\n\n