AMD's FSR Redstone Upscaling to Support Nvidia GPUs

AMD is paving the way for a more inclusive graphics ecosystem with its forthcoming FSR Redstone upscaling technology. This innovation promises broad compatibility, extending its benefits beyond AMD's own hardware to include rival GPUs like those from Nvidia. This move highlights AMD's strategic vision for leveraging AI in graphics processing, focusing on software-driven solutions that could redefine industry standards for performance and accessibility.

AMD's FSR Redstone Upscaling: Bridging GPU Divides and Advancing AI Integration

In a significant disclosure during an interview with the Japanese publication 4Gamer, Chris Hall, Senior Director of Software Development and head of AMD's ROCm project, revealed groundbreaking details about FSR Redstone. This next iteration of AMD's FidelityFX Super Resolution, possibly to be branded as FSR 5, is being developed with an emphasis on universal compatibility. The core of this advancement lies in AMD's ML2CODE (Machine Learning to Code) framework, a component of the ROCm project, which is AMD's answer to Nvidia's CUDA for AI development.

According to Hall, FSR Redstone's neural rendering technology translates into compute shader code. This ingenious approach means that the neural rendering capabilities of FSR Redstone are not confined to AMD hardware but can execute on graphics processors from other companies, notably Nvidia. This is a crucial distinction from Nvidia's DLSS technology, which necessitates specialized Tensor cores for AI processing. FSR Redstone, in contrast, harnesses conventional GPU shaders, sidestepping the requirement for dedicated AI or matrix math cores.

Hall elaborated that ML2CODE-based frameworks, such as FSR Redstone, offer seamless integration directly into DirectX or Vulkan graphics pipelines, ensuring minimal latency. He posited that the ML2CODE solution represents the most effective current method for harmonizing 3D graphics and AI technologies. This philosophy underscores AMD's preference for employing AI-enhanced upscaling via the shader pipeline, a divergence from Nvidia's long-standing strategy of incorporating dedicated AI cores across multiple generations of its RTX GPUs.

Despite this current focus, Hall also indicated that future AMD graphics architectures might integrate dedicated AI cores, specifically designed for compatibility with Microsoft's DirectX Cooperative Vector technology. While acknowledging the potential for significant delays if a GPU lacks an architecture optimized for Cooperative Vector technology, Hall affirmed AMD's active engagement in supporting this framework. He envisions a future where the two methodologies—shader pipeline-based and dedicated AI core-based ML upscaling—might converge or be integrated, though the timeline for such a fusion remains uncertain.

This commitment to open standards is consistent with AMD's current GPU technology. Even the latest RDNA 4 GPUs, such as the Radeon RX 9070 XT, feature shader engine optimizations for improved matrix math performance, which AMD refers to as an "AI accelerator," rather than discrete hardware AI cores. Hall emphasized the versatility and optimal performance of the ML2CODE-based Compute Shader approach across a wide array of existing GPUs, including those from Nvidia.

While the technical feasibility for FSR Redstone to operate on Nvidia and Intel GPUs appears robust, the ultimate decision regarding cross-platform availability will likely hinge on commercial and political considerations. For gamers, the prospect of platform-agnostic upscaling and frame generation technologies is highly appealing, as it would empower them to select the most effective upscaler for any given game, irrespective of their GPU vendor. This could also mean that older AMD GPUs, such as the RDNA 3-based RX 7000 series, including the RX 7800 XT, might gain compatibility with FSR Redstone.

Conversely, Nvidia's DLSS platform remains proprietary, exclusively supported on its own GPUs. This proprietary stance gives Nvidia a strategic advantage, as GPUs that support both major vendors' upscaling technologies will, for the foreseeable future, likely be Nvidia's offerings. This creates a fascinating competitive dynamic in the rapidly evolving landscape of graphics technology.

Embracing Openness: A Catalyst for Innovation and Consumer Choice in Graphics Technology

AMD's strategic move to broaden the compatibility of its FSR Redstone upscaling technology signifies a pivotal moment in the graphics card industry. By designing a solution that transcends hardware boundaries, AMD is championing an open ecosystem, fostering competition, and ultimately benefiting the end-user. This approach encourages a future where technological advancements prioritize accessibility and choice over proprietary lock-ins. It challenges the established norms and could inspire other industry players to reconsider their strategies, driving innovation towards universally beneficial standards. For consumers, this translates into greater flexibility and potentially superior gaming experiences, regardless of their chosen hardware vendor, marking a positive shift towards a more collaborative and user-centric technological landscape.