What is XeSS? Understanding Intel’s Oversampling Technology to Boost Gaming


Today Intel announced a number of new products and technologies in and related to its Arc line of discrete mobile graphics cards. While you can read the full suite of announcements at the previous link, in this article we’ll discuss Intel’s foray into the world of AI-assisted image scaling, known as name of “Intel XeSS”.

And what a melee is that incursion these days! Intel’s new technology enters a rapidly changing graphics market that sees rival graphics gamers Nvidia and AMD competing with the same technologies to get the most out of many of their existing graphics solutions. At a time when GPUs are rare and expensive, this kind of efficiency technology is increasingly important. Here’s a quick look at the new technology, which we haven’t had a chance to try yet.

What does XeSS mean?

XeSS is short for “Xe Supersampling”. (The “Xe” is Intel’s generic name for its latest graphics solutions.)

What do supersamplers and upscalers like XeSS do?

Intel’s XeSS will work to do the same as Nvidia’s Deep-Learning Supersampling (DLSS, now in version 2.3) and AMD‘s dynamic technology duo of FidelityFX Super Resolution (FSR) and Radeon Super Resolution (RSR). The idea is to give you higher frame rates in your favorite games on the same hardware, using AI-assisted training and algorithms to handle the task.

At these three links, we have much deeper dives into how these specific technologies work their craft. There are specific nuances and differences that set them apart, as well as guidelines for the hardware and games they will work with. It is a tangled field. Some technologies work on many games with very specific GPUs, and vice versa.

How do these technologies work?

This kind of frame rate efficiency is achieved by lowering the game’s rendering resolution (which decreases the load on the GPU, allowing for higher frame rates) and then “oversampling” (also known as ” upscaling”, depending on the technique), render the image to a point where your eyes cannot see the difference in quality. “More frames for free” is the colloquial phrase around these technologies these days.

We tested all three existing players – again, check out the links above for those deep dives – and found that Nvidia’s DLSS consistently produced the best image quality. However, DLSS gaming compatibility is limited (more on that in a moment), and it only works with Nvidia GeForce RTX cards.

Intel’s XeSS technology is trained more similarly to Nvidia’s DLSS than AMD’s RSR. This is a good sign for the possible levels of quality we might expect from XeSS in various PC games when it launches. However, there is a catch, as is often the case with this type of technology: the number of games at launch that can use it.

What games does XeSS work with?

All of this excitement around a new supersampler/upscaler entering the arena is unfortunately tempered by one major downside: the fact that per-game training is required to deploy it.

Like DLSS, XeSS will require every game it works on to be trained before release (at Intel’s Arc labs, in the case of XeSS). The training acclimatizes the technology to gaming scenes and basically explains how it can cut corners without losing quality. This aspect of training, however, could be a potentially significant barrier to short-term adoption of XeSS.

List of Intel XeSS games

Presenting us with this XeSS detail during a pre-launch briefing, Intel presenters said that while the company explores “all forms of neural scaling” at launch, XeSS will only have 14 compatible games. . Here is the full list of titles ready to launch:

By comparison, Nvidia’s DLSS is approaching the 180 title mark, while AMD’s RSR is compatible with thousands of games, almost any DirectX 11 or DirectX 12/12+ based title. To forgive the pun, initially there will be no excess XeSS.

What kind of horizon could we be talking about here for Intel? For reference: it took over three years for the library of DLSS-trained games to reach its current number. And many of these DLSS-enabled games have long been in and out of the gamers’ zeitgeist.

What hardware will XeSS run on?

Intel’s Arc discrete graphics chips will be the first hardware to support XeSS with the list of supported games. Arc will debut first in a series of laptops (definitely, not hardcore gaming models) using the Arc 3 line of mobile Intel Arc GPUs; the first laptops will be rolled out in April. (Intel has detailed two Arc 3 GPUs.)

Arc 5 and 7 chips for laptops are expected to start appearing in laptops in early summer 2022, with desktop Arc expansion cards to follow later. (Intel has yet to announce a firm timeline for these.)

When will Intel XeSS be launched?

Behind Nvidia, Intel will be the second graphics player to integrate AI-backed rendering cores into its discrete mobile solutions, something AMD has yet to achieve. This gives the company an edge over a launch company; however, all versions of Nvidia’s GeForce RTX 30 Series mobile chips come with DLSS compatibility, driven by their built-in Tensor cores. Intel says XeSS compatibility will extend to both mobile and discrete graphics solutions, as well as desktop SoCs, at some point in the future.

Intel Xe Super Sampling

Although we got a glimpse of Intel’s XeSS during our pre-briefing, availability information and launch dates outside of “early summer 2022” are still under wraps. Stay tuned as we test the first batch of titles… whenever Intel’s solid launch date for XeSS arrives.

What's New Now to get our top stories delivered to your inbox every morning.","first_published_at":"2021-09-30T21:30:40.000000Z","published_at":"2022-03-29T17:10:02.000000Z","last_published_at":"2022-03-29T17:09:22.000000Z","created_at":null,"updated_at":"2022-03-29T17:10:02.000000Z"})" x-show="showEmailSignUp()" class="rounded bg-gray-lightest text-center md:px-32 md:py-8 p-4 mt-8 container-xs">
Receive our best stories!

Sign up for What’s up now to get our top stories delivered to your inbox every morning.

This newsletter may contain advertisements, offers or affiliate links. Signing up for a newsletter indicates your consent to our Terms of Use and Privacy Policy. You can unsubscribe from newsletters at any time.


Comments are closed.