Skip to main content
  1. Home
  2. Computing
  3. News

AMD Radeon VII will support DLSS-like upscaling developed by Microsoft

Add as a preferred source on Google
Image used with permission by copyright holder

AMD’s Radeon VII may be more than just a direct performance competitor for Nvidia’s RTX 2080, it might be able to offer similar features, too. Although AMD has ruled out ray tracing on its graphics cards until it becomes more of a mainstream feature, it could adopt something akin to Nvidia’s deep learning super sampling (DLSS). Its plan is to support an open alternative developed by Microsoft called DirectML.

One of the big selling points for Nvidia’s latest Turing-generation of graphics cards was its feature set. Beyond the raw power of the new GPUs, Nvidia talked up their ability to handle ray tracing and DLSS, the latter of which is a form of upscaling and anti-aliasing that is designed to produce 4K-like visuals, at a 1,440p overhead. It was a bespoke technology, as far as Nvidia was concerned, enabled only through its specifically designed Tensor cores which populate every RTX graphics card.

Recommended Videos

But AMD is looking to go a different route to enable much the same technology but power it with its standard GPU cores. Microsoft’s DirectML was announced back in March 2018, with Microsoft talking up the potential of bringing deep neural networks (DNN) to gaming. They could be used for everything from on-the-fly difficulty adjustments in-game, to streamlining the developmental process. But it also discussed the potential to use DirectML for supersampling, and even gives an example of Nvidia doing so.

AMD’s Adam Kozak conducted an interview with Japanese tech site, 4Gamer (via Toms Hardware) and suggested that AMD’s upcoming 7nm Vega 2 Radeon VII graphics card was fully verified to utilize the technology and that early test results with the card showed positive results.

As exciting as it is for gamers that such a feature would be available on AMD hardware, it’s even more impactful that this is a technology that isn’t locked down to a specific manufacturer. In much the same way that AMD’s Freesync technology is more affordable and widely available than Nvidia’s GSync, DirectML could have a much lower barrier for entry. That goes doubly so when you consider that AMD graphics are expected to power the next-generation of game consoles. If the company adopts DirectML super sampling, then it seems likely that game developers would do so en masse in turn.

The only hitch to this new technology is that it requires DirectX12. Although there are still a limited number of DX12-supporting games at this time, that will likely change in the years to come, which could make DirectML an important graphics technology in turn.

Jon Martindale
Jon Martindale covers how to guides, best-of lists, and explainers to help everyone understand the hottest new hardware and…
Canva Code 2.0 just made vibe coding way less intimidating for everyone
Canva Code 2.0 feature

Coding used to be reserved for developers who spent years learning complex languages. That has slowly changed with vibe coding, which lets you build apps and websites using simple, plain-language prompts. 

The problem is that most of these tools still feel intimidating for regular folks, as they still need to understand the code to make any meaningful changes. If not, everything you make tends to look the same.

Read more
Windows users can finally pick when updates stop with Microsoft’s latest patch
From pausing updates on your own schedule to rolling back a broken PC in one click, here's everything new in Windows 11's July 2026 update.
Windows 11 Laptop

Patch Tuesday updates are usually a shrug-and-install affair, but Microsoft's July 2026 release actually gives you something to be excited about.

You can grab this update, tagged KB5101650, right now through Settings, or manually via the Microsoft Update Catalog if you'd rather not wait for it to roll out.

Read more
Can AI audiobooks narrate better than humans? This study says many listeners think so
New study finds listeners favor AI narrated audiobooks over traditional human narration in blind testing.
Audiobooks on Spotify on an iPhone.

You might assume most listeners would pick a real human voice over a synthetic one, but a new study says otherwise. Edison Research at SSRS surveyed 1,005 fiction audiobook fans in May 2026 for a study commissioned by AI audio company Spoken. The twist is that listeners rated the AI narration higher, and they did not even know it was AI until after they heard it (via Variety).

Why listeners favored the AI narration

Read more