The Mother Fucker 26 Posted October 16, 2019 Share Posted October 16, 2019 1 minute ago, Remij_ said: That's not what I think.. it's a fact that the game uses them. What we're trying to figure out is how much or little it can make a difference... It would help.... if you actually ran at the same settings as we did. Since I can't run the same resolution as you, I'm better position with 4 desktops and various similar setups to test the theory, and so far it does not look like CPU is making much a difference. IMO, it looks like the Turing GPU is well optimized for RT than the rest. Quote Link to post Share on other sites
Remij 4,683 Posted October 16, 2019 Author Share Posted October 16, 2019 5 hours ago, lynux3 said: I can, he blew all his savings on a card that'll be outdated in less than 2 years. blew all my savings and I upgrade every year Quote Link to post Share on other sites
Remij 4,683 Posted October 16, 2019 Author Share Posted October 16, 2019 6 minutes ago, The Mother Fucker said: Since I can't run the same resolution as you, I'm better position with 4 desktops and various similar setups to test the theory, and so far it does not look like CPU is making much a difference. IMO, it looks like the Turing GPU is well optimized for RT than the rest. Well, I guess the CPU is only going to help you to a point.. then you're pretty much GPU bound regardless. As I said though.. this is a very limited in scope implementation. It's only being used for the shadows of the tanks.. nothing else. It's much easier on the hardware because they know the position of the light source (the sun) and only have to trace pixels from the ground around the tank back to the light source. It's extremely simplistic, compared to the things Nvidia and RTX developers have been working on. That's not a knock against it.. it's just a fact. There's no light bouncing here.. like Metro's global illumination which needs to calculate rays bouncing across ALL objects in the scene.. not just some. And Metro has tons of dynamic objects in the scene.. and not to mention it looks far better in general. This demo has a single light source casting shadows. It's similar in the way it works to that Crytek RT demo they made. That only used mirror reflections... any other diffused surface would not reflect anything. A mirror reflection is easy because you know exactly where the ray is travelling. A diffused or rough surface would cause the rays to have to be bounced in all sorts of directions and doing that math is far harder. It's still very impressive.. but Metro's DLC for example has multiple dynamic emissive lights of all colors casting rays out into the scene coloring, shading, blending and bouncing. It's FAR more technically demanding.. and you can begin to see why they can't use the CPU for that and why GPU acceleration is necessary. It's why Control's RT is so impressive. It's using multiple effects all at once and performance is still really good considering that. RT Shadows, RT Ambient Occlusion, RT Reflections, RT Indirect Diffused lighting... ect. Quote Link to post Share on other sites
The Mother Fucker 26 Posted October 17, 2019 Share Posted October 17, 2019 I meant to say this test demo is well optimized for the Turing GPU. We need to see more results. @JONBpc run this demo in 1080p with RT off, and then run it again with RT on. I'm curious to see what you get. Quote Link to post Share on other sites
JonDnD 2,644 Posted October 17, 2019 Share Posted October 17, 2019 Maybe sometime tomorrow Quote Link to post Share on other sites
kaz 2,439 Posted October 17, 2019 Share Posted October 17, 2019 benchmarking world of tanks fucking hermsmells Quote Link to post Share on other sites
JonDnD 2,644 Posted October 18, 2019 Share Posted October 18, 2019 (edited) Ray Tracing off I got 29585 ultra Ray Tracing on I got 19482 ultra Edited October 18, 2019 by JONBpc Quote Link to post Share on other sites
The Mother Fucker 26 Posted October 18, 2019 Share Posted October 18, 2019 (edited) 34 minutes ago, JONBpc said: Ray Tracing off I got 29585 ultra Ray Tracing on I got 19482 ultra I wonder why RT on/off made a big difference with your GPU than mine. My GTX 1660Ti Edited October 18, 2019 by The Mother Fucker Quote Link to post Share on other sites
Remij 4,683 Posted October 18, 2019 Author Share Posted October 18, 2019 Just now, The Mother Fucker said: I wonder why RT on/off made a big difference with your GPU than mine. Maybe because he's doing Ultra and you're doing High? Quote Link to post Share on other sites
The Mother Fucker 26 Posted October 18, 2019 Share Posted October 18, 2019 Just now, Remij_ said: Maybe because he's doing Ultra and you're doing High? are you blind? I did a benchmark for every scenario. Quote Link to post Share on other sites
Remij 4,683 Posted October 18, 2019 Author Share Posted October 18, 2019 (edited) 3 minutes ago, The Mother Fucker said: are you blind? I did a benchmark for every scenario. Whatever. who could tell? You posted a bunch of bullshit from 4 different setups... Anyway... his hit from RT isn't as big as yours was. You're ~35K RT off, and ~20K RT on. He's ~30K RT off, and ~20K RT on. Since you edited to mean your 1660ti... that's because you're GPU limited. Edited October 18, 2019 by Remij_ 1 Quote Link to post Share on other sites
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.