Search found 164 matches

by graphicsMan
Thu Oct 04, 2012 6:14 pm
Forum: Visuals, Tools, Demos & Sources
Topic: Siliconarts Announces 'RayCore' Real-Time Ray Tracing GPU
Replies: 3
Views: 4709

Re: Siliconarts Announces 'RayCore' Real-Time Ray Tracing GP

I guess the difference between this and caustic graphics is that this is presumably tied to a fixed tracing/shading pipeline?

"considered impossible to implement in hardware" by whom exactly?

sigh
by graphicsMan
Wed Oct 03, 2012 6:11 pm
Forum: Visuals, Tools, Demos & Sources
Topic: Parallella: A Supercomputer For Everyone
Replies: 5
Views: 5791

Re: Parallella: A Supercomputer For Everyone

:) Yes, with a RAM size of 1GB for a ton of MIMD cores, I somehow doubt that for our use case it will scale easily ;)
by graphicsMan
Tue Oct 02, 2012 9:59 pm
Forum: Visuals, Tools, Demos & Sources
Topic: Parallella: A Supercomputer For Everyone
Replies: 5
Views: 5791

Re: Parallella: A Supercomputer For Everyone

The programming difficulty for anything non-trivial should probably be considered for both of the high performance options :) Also, achieving close to max perf on an Intel chip is a lot easier than getting close to max perf on a GPU. Unsure about the parallella machine.
by graphicsMan
Tue Sep 11, 2012 7:24 pm
Forum: General Development
Topic: Quadrics kinda sorta work for real time
Replies: 34
Views: 24202

Re: Quadrics kinda sorta work for real time

I'm looking forward to the point where you can provide memory/speed comparisons between small tessellation into quadrics vs deep tessellation into triangles at the same perceptual quality :) I will be eagerly watching this thread ;)
by graphicsMan
Tue Sep 11, 2012 4:16 pm
Forum: General Development
Topic: Quadrics kinda sorta work for real time
Replies: 34
Views: 24202

Re: Quadrics kinda sorta work for real time

cool. Can you explain more? :) Which what is kinda working?
by graphicsMan
Fri Sep 07, 2012 6:23 pm
Forum: General Development
Topic: Nested spatial data structures
Replies: 4
Views: 3813

Re: Nested spatial data structures

Yeah, these types of approaches have been used for a very long time actually. Matt Pharr used something like this for his out-of-core ray tracing work in the 90s. I suspect the idea is even several years older than that. Not to burst your bubble or anything though :) I wouldn't by any stretch say th...
by graphicsMan
Tue Sep 04, 2012 1:15 am
Forum: General Development
Topic: sampling multiple light source
Replies: 7
Views: 4554

Re: sampling multiple light source

You have this: pdf = [0.2, 0.2, 0.6] cdf = [0.2, 0.4, 1.0] If you have a random value of 0.55, you perform a binary search to find the bin with cdf value greater to or equal to your random value. Therefore, it will put you at (0-based) index of two. No warping is necessary here. If you have an unnor...
by graphicsMan
Mon Sep 03, 2012 3:53 pm
Forum: General Development
Topic: sampling multiple light source
Replies: 7
Views: 4554

Re: sampling multiple light source

Right, with only three lights and those probabilities, if you don't want to create a cdf explicitly and binary search it (which is what you'll do when you want more light sources...), you can do: if(p < 0.2) { choose light 0 power /= 0.2; } else if (p < 0.4) { choose light 1 power /= 0.2; } else { c...
by graphicsMan
Mon Sep 03, 2012 12:59 am
Forum: General Development
Topic: sampling multiple light source
Replies: 7
Views: 4554

Re: sampling multiple light source

Your probabilities need to add up to one. That could mean in your case that you select the sky or no light at all with probability of 0.034. If you will always select one of the three lights, you need to normalize your probabilities (divide each probability by the sum of the probabilities), create a...
by graphicsMan
Fri Aug 17, 2012 4:16 am
Forum: General Development
Topic: Anybody thinks seriously about diffraction in RT/PT?
Replies: 7
Views: 7152

Re: Anybody thinks seriously about diffraction in RT/PT?

It's interesting, but not the same kind of simulation. It's more of a micro-facet level interaction instead of macro geometry interaction.