> However Nvidia are treating DLSS as their secret sauce and not publishing any details

https://developer.nvidia.com/gtc/2020/video/s22698

There are no real details in that presentation, they explain what input data DLSS needs, and demonstrate the outputs it can produce, but gloss over how it actually gets there.

From the inputs you can glean that DLSS is using temporal integration but that's hardly a new idea, the novel part is in how it performs the integration and Nvidia isn't sharing that.

It is using TAA-upsampling and I think the only DL part is replacing the neighborhood clamping filter with a trained function.

You can look at the tradeoffs of neighborhood clamping/clipping here:

https://de45xmedrsdbp.cloudfront.net/Resources/files/Tempora...

Those slides were before TAA upsampling was added, but the issues are similar once it is in place too. Unreal Engine already has TAA upsampling, but has to fight against the issues mentioned in the Nvidia presentation.

Nvidia trained it for specific upsampling multiples, so the main functional difference with existing TAA upsampling is it can't do continuous base-resolution changes.

It's basically just a good TAA and the tensor cores are basically used as a poor man's Lanczos. The tensor cores are used for obfuscation, and to make sure it doesn't run on AMD. Very little about DLSS 2.0 is "DL".

This is inaccurate.

Please be aware that the parent is Brian Catanzaro, VP of Applied Deep Learning Research at NVidia, who overviews the development of DLSS among many other things.

If he says that the GP is inaccurate, trust him.

Hi, Brian!

This appears to be more accurate than DLSS according to Fig. 11 in the paper, and has the 16x mode as well as 4x.

Meanwhile DLSS is real time and this takes 100ms a frame.

DLSS is possibly baked into the graphics cards. It is what the RT cores are for. Facebook is just releasing a POC. I doubt the code is handwritten CUDA optimized for performance.

On Nvidia cards the RT cores are the ones handling BVH traversal and ray-triangle intersection operations.

The part of the hardware that is running the DLSS ML model are the tensor cores. But the algorithm and the model itself is not baked in, it is provided in the driver and/or game

That was DLSS 1.0 they compared with.