AMD FidelityFX Super Resolution to Introduce AI-Based Upscaling

AMD plans for FSR4 to be fully AI-based — designed to improve quality and maximize power efficiency

The team in charge has already been working on this for 9–12 months.

[URL unfurl="true"]https://www.tomshardware.com/pc-components/gpus/amd-plans-for-fsr4-to-be-fully-ai-based-designed-to-improve-quality-and-maximize-power-efficiency[/URL]

Nice. They say AI. I say api.

"Grimlakin, post: 89373, member: 215" wrote:

Nice. They say AI. I say api.

For those who are more into the coding side of things the Tom's article does have some interesting theories about how it will happen along with a little explanation with how Intel and NVIDIA are doing it. Its towards the end of their article.

FSR 4 Redstone Pathtracing (Firestone) could run on 'any' GPU thanks to ML2CODE framework

"Marees, post: 97994, member: 1536" wrote:

AMD announces FSR Redstone for RDNA4: Neural Radiance Caching, ML Ray Regeneration and Frame Generation



"The next step is FSR Redstone, a new version of super resolution and frame generation. This version will focus on three key features: Neural Radiance Caching, machine learning-accelerated Ray Regeneration, and machine learning accelerated Frame Generation."





"Hall began by saying something surprising. FSR Redstone is implemented so that it can run even if the GPU does not have AI acceleration capabilities at runtime. Despite being a machine learning/AI-based technology, it does not require AI acceleration on the GPU side."





"FSR Redstone was developed using AMD ML2CODE (Machine Learning to Code), a research project from ROCm. The core part of the neural rendering technology is converted into optimized Compute Shader code by utilizing ML2CODE."





ML2CODE is included in AMD's GPU computing (GPGPU) platform, ROCm 6.1 and later.



Rather than executing trained AI cores at runtime, it optimizes them as existing Compute Shader code and enables native execution.





[embedded media]

A Japanese tech journalist, Zenji Nishikawa, came to San Jose, USA & interviewed Chris Hall AMD's software head of ROCm on FSR 4 Firestone

It appears Firestone path tracing software can run on any GPU since it is compiled to shader code using ML2CODE framework

[embedded media]

"Marees, post: 98009, member: 1536" wrote:

It appears Firestone path tracing software can run on any GPU since it is compiled to shader code using ML2CODE framework

(see other such attempts)

*press X to doubt*

I'll believe it works well when I see performance comparisons

What... are you suggesting emulation 9 x out of 10 is slower than native?

"Tsing, post: 82927, member: 5" wrote:

A future version of AMD FidelityFX Super Resolution will leverage AI as part of its upscaling pipeline...

"DrezKill, post: 98035, member: 230" wrote:

Wait, a "future version"? Is it not that already? I thought the whole point of FSR4 to begin with was that it was AI-based?

"DrezKill, post: 98035, member: 230" wrote:

Wait, a "future version"? Is it not that already? I thought the whole point of FSR4 to begin with was that it was AI-based?

Future is here — FSR 4 hacked to work on RDNA 3 & 2 GPUs

[URL unfurl="true"]https://hardforum.com/threads/amd-fsr-4-source-code-leaked-probably-hackable-for-older-gpus.2043060/[/URL]

"Marees, post: 98073, member: 1536" wrote:

Did you check the date 😀

Ray Regeneration is now live — on RDNA 4 hardware

[embedded media]

"Marees, post: 98009, member: 1536" wrote:

FSR 4 Redstone Pathtracing (Firestone) could run on 'any' GPU thanks to ML2CODE framework







A Japanese tech journalist, Zenji Nishikawa, came to San Jose, USA & interviewed Chris Hall AMD's software head of ROCm on FSR 4 Firestone



It appears Firestone path tracing software can run on any GPU since it is compiled to shader code using ML2CODE framework





[embedded media]

[embedded media]