Suggestions regarding how RTX tech can be helpful in DCS

I know we talk a lot about RTX and DCS, but many of them goes into pricing, risk/uncertainty and AMD etc.

How about we discuss about how RTX can be helpful in DCS, just assuming that raytracing tech will be there sooner or later?

What i am guessing is:

1. Terrain and even exterior rendering looks hard to utilize RTX tech since in DCS it's so open and wide, no need to "individually trace" some solar rays within DCS world. But i dont know better.

2. Maybe cockpit rendering and lighting can utilize RTX tech. This way usual CUDA cores can focus on rendering usual things, RTX/Tensor cores can work on internal cockpit rendering, at least some of them, to better utilize GPU power and functions.

3. Maybe other ways to utilize tensor core for other than rendering? like increasing view distance? Combined Arms AI? (if possible i think after year 2030 :)

Just writing in fear that right now RTX Tensor cores being useless in DCS. Hope we can pool some ideas and maybe DCS can use the things that new GPUs provide soon enough..

All that stuff requires Dx12 afaik. DCS is about to switch to Vulkan, so the pros and cons of Dx12 tech are purely academic. If any of that stuff can be utilised without Dx12 it may make its way in someday, but I'm pretty sure it's all proprietary.

A flight sim would be ideal for ray-tracing:

Shadows can be accurately processed through RTX and the DLSS can help significantly with getting better image quality with less GPU use.

In fact, raytracing makes lighting so much simpler: you have either a sun or a moon. You have clouds. You have atmospheric effects on the light. You have materials that interact with the light.

Compare that with traditional GPU processing: prebaking everything which gives everything a nondynamic lighting. Prebaking requires considerable resources of an artwork team that can't be altered easily as you can with raytracing.

What about DLSS?

I lot of talk about ray tracing which would be awesome, of course. The Vulcan API does support it as well AFAIK. Lighting that is ray traced or simulated with another technique is looking more expensive than antialiasing, add in development time, driver support, etc. and the next generation of cards might be out before we see it used well in DCS. DLSS on the other hand,...

I watched the keynote, from what I gathered DLSS will allow a game or sim to be rendered at say 1080p or 1440p but supersampled up to 4K+ with minimal hit on CUDA cores. It uses the tensor cores to apply machine learned algorithms to imagery. By a process of example, trial, and error an AI neural net created supersampled images that could fool the human eye but with much less processing power needed. (i.e. don't upscale portions of screen with solid colors). Brute force supersampling is very expensive. If it could do this for say the resources of MSAA techniques we could see some big resolution gains over the next couple years. Maybe think of the tensor cores as an ASIC for supersampling like TVs upscale to 4K. Not as good as native 4K, but still looks better than a 1080p blu ray. 8K displays will start to be mass produced by next year.

How this is exactly applied to currently existing games, I have no idea. Would the neural net need to play DCS for a couple weeks and then release the driver? Input lag? All still current unknowns.

If it's something DCS can do without expending too many resources, awesome. People like eye candy and DCS eye candy is really good. It could really grow the player base. Some high rez YT videos with RTX lighting could have a nice wow factor for noobs to flight sims. Help make people take the leap from console to PC or buy that HOTAS.