The battle for supremacy in the GPU market is heating up, with AMD research suggesting plans to challenge Nvidia’s dominance using neural supersampling and denoising for real-time path tracing. While Nvidia has been leading the charge with its advanced AI and machine learning-based technologies, AMD is gearing up to close the gap.
A New Approach: Neural Supersampling and Denoising
AMD’s latest research focuses on achieving real-time path tracing on RDNA GPUs via neural network solutions. This approach is designed to address the limitations of traditional rendering methods, which require thousands or even tens of thousands of ray calculations per pixel. In contrast, AMD’s neural supersampling and denoising method can generate high-quality images with a significantly reduced number of samples.
The Problem with Path Tracing
Path tracing is considered the gold standard for rendering graphics, as it allows for accurate simulation of light behavior in a scene. However, this process requires an enormous amount of computational power, making it impractical for real-time applications. To overcome this limitation, AMD is using neural networks to denoise and upsample images, effectively reconstructing scene details with a few samples per pixel.
The Workflow: A Single Neural Network
AMD’s innovative approach combines upscaling and denoising within a single neural network. This unified process can replace multiple traditional rendering algorithms and upscalers, making it more efficient and effective. The result is a high-quality image that meets the standards of Nvidia’s DLSS technologies.
The Potential Impact: A New Version of FSR
This research could potentially lead to a new version of AMD’s FidelityFX Super Resolution (FSR) technology, which might match Nvidia’s performance and image quality standards. With FSR 2.0, gamers can enjoy high-quality graphics on lower-end hardware, but the limitations of current GPUs might restrict the quality and features of this new iteration.
The Challenges Ahead: Hardware Requirements
While AMD is working to improve its neural supersampling and denoising method, there are still significant challenges ahead. The company will need to determine whether it can run these complex algorithms on existing hardware or if dedicated AI acceleration features are required. If the latter is true, this might limit the quality and performance of FSR 2.0.
The Road Ahead: A Brighter Future for Gamers
If AMD succeeds in delivering a refined approach to neural path tracing and upscaling, it could bring high-fidelity graphics to a broader range of hardware. This would be a significant step forward for gamers, who can look forward to enjoying more realistic and immersive experiences on lower-end hardware.