AMD’s Radeon VII Supports DirectML – An Alternative to DLSS?
AMD’s Radeon VII Supports DirectML – An Alternative to DLSS?
Think of DirectML as the Machine Learning equivalent of DXR (DirectX Raytracing), allowing DirectX 12 to support advanced features and utilise AI to improve future games. Â
An example of how AI can be seen in Nvidia’s DLSS (Deep Learning Super Sampling) technology, which uses a Deep Learning algorithm to boost game performance while offering a final image that can provide superior image quality to TAA (Temporal Anti Aliasing). In the past, Microsoft has showcased DirectML achieving similar feats, which means that there may soon be a multi-vendor alternative to Nvidia’s DLSS technology.  Â
DirectML supports all DirectX 12 compliant hardware, much like DXR, and like DXR it can also exploit the hardware acceleration capabilities of modern graphics architectures. In effect, this will allow developers to access hardware features like Nvidia’s Tensor cores, just like how DXR enables developers to utilise Turing’s RT cores. In the case of DirectML, the performance of AMD’s Radeon VII could be used to deliver a “DLSS-like” effect, but using an approach that will work on Radeon hardware.
AMD’s Adam Kozak stated that the (translated) “Radeon VII shows excellent results” when the company experimented with DirectML.  Â
Â
Â
Microsoft has already showcased the potential of machine learning in gaming applications, with the image below showcasing what happens when Machine Learning is used to upscale an image to four times its original resolution (basically from 1080p to 4K) to generate a sharper final image and reduced aliasing. The image below is a comparison between ML Super Sampling and bilinear upsampling.Â
This technique has also been showcased during one of Microsoft’s SIGGRAPH 2018 tech talks. This talk, which is entitled “Deep Learning for Real-Time Rendering: Accelerating GPU Inferencing with DirectML and DirectX 12” showcases Nvidia hardware upscaling Playground Games’ Forza Horizon 3 from 1080p to 4K using DirectML in real time. DirectML has the potential to improve the graphical fidelity of future console and PC games. Â
Â
   We couldn’t write a graphics blog without calling out how DNNs can help improve the visual quality and performance of games. Take a close look at what happens when NVIDIA uses ML to up-sample this photo of a car by 4x. At first the images will look quite similar, but when you zoom in close, you’ll notice that the car on the right has some jagged edges, or aliasing, and the one using ML on the left is crisper. Models can learn to determine the best color for each pixel to benefit small images that are upscaled, or images that are zoomed in on. You may have had the experience when playing a game where objects look great from afar, but when you move close to a wall or hide behind a crate, things start to look a bit blocky or fuzzy – with ML we may see the end of those types of experiences. Â
With DXR and DirectML, it is clear that gaming will become more than just a game of pure shading performance. AMD, Nvidia and Intel will all need to develop hardware that is capable of handling inferencing, Ray Tracing and traditional shading, opening up a new hardware arms race that will transform gaming over time.
With the Radeon VII being DirectML compliant, AMD has taken a step into the world of AI in gaming, giving them the opportunity to develop an alternative to Nvidia’s DLSS technology, one which as the potential to impact users outside of the Nvidia GPU ecosystem. If successful, an AMD-powered AI upsampling has a lot of potential for use within both next-generation games consoles and PCs users alike.Â
DirectML support is coming to Windows 10 as part of its Spring 2019 update. Expect to hear more about DirectML at GDC 2019.Â
You can join the discussion on AMD’s Radeon VII graphics card supporting DirectML on the OC3D Forums.Â