Nvidia Blackwell Architecture Speculation

  • Thread starter Deleted member 2197
  • Start date
alking DLSS TNN Performance now.
The difference in motion is indeed dramatic. Below are a few comparisons taken in CP2077 during fast forward motion (hence the minor mistmatch). One screenshot in each pair is a native resolution with TAA, while the other is DLSS Performance (with the default sharpening level for DLSS). Display resolution is 1440p, meaning the rendering resolution with DLSS is 1280x720. I intentionally did not label the screenshots so that anyone interested can try to identify DLSS on their own.
1. 2. 3. 4.
 
DLSS and Native TAA can be swapped in position for each pair.
Really easy to spot DLSS, even without zooming. Just swipe couple times to find something that is 'different' and the nicer/sharper one is DLSS. Trees in mid distance. Or the trucks LED array lights. Last one just everything at a distance is clearer with DLSS.

BUT another way to find DLSS... zoom a bit and look for odd stippling/checkering. Its like everywhere with DLSS, eg:
Weird how common it is while image is also clearly better.
 
this video is one of the best examples of native vs DLSS 4, and it looks much better overall in DLSS 4, I didn't expect those differences tbh.

Especially the HUGE difference in antialiasing when he compares Stalker, which is a jaggies fest in native-.

The guy does a good job comparing both in real time without actually splitting the screen but superposing.

 
Nvidia Blackwell Architecture Whitepaper.
https://images.nvidia.com/aem-dam/S...ell/nvidia-rtx-blackwell-gpu-architecture.pdf

Mega geometry details are quite thought provoking.

DXR 2.0 must be coming. This is a pretty significant departure from how ray-tracing is currently being handled.

I'm really curious how Mega Geometry works on a gpu that lacks the cluster intersection testing in the RT core. You'll still get big benefits on the CPU side, but I wonder how the intersection testing is handled and what the performance penalty is.
 
DXR 2.0 must be coming. This is a pretty significant departure from how ray-tracing is currently being handled.

I'm really curious how Mega Geometry works on a gpu that lacks the cluster intersection testing in the RT core. You'll still get big benefits on the CPU side, but I wonder how the intersection testing is handled and what the performance penalty is.
Based on AW2 - RTX 2080 Ti Runs about 5 ms better with it on vs off
 
Guess nobody is really interested in talking about the 5080 reviews, eh? lol


Turns out we didn't need to wait for reviews to know this was gonna be extremely disappointing.

Honestly, even Blackwell as an architecture is super underwhelming. You can argue it's setting up some improvements that will show up later down the line, but by the time these things are more commonly incorporated into games, there will be better GPU's out.
 
Back
Top