Search found 28 matches

by ziu
Tue Nov 05, 2013 2:20 pm
Forum: General Development
Topic: Is it still worth it to do CPU SIMD ray tracing?
Replies: 12
Views: 8967

Re: Is it still worth it to do CPU SIMD ray tracing?

In my opinion ignore SSE intrinsics and use either OpenCL or OpenACC. That way your code will be accelerated on both CPUs and GPUs. Who knows you may even want to port from x86 to ARM someday. Or when Intel changes AVX to something else again for that matter.
by ziu
Sun Oct 20, 2013 9:50 pm
Forum: Links & Papers
Topic: Improving Data Locality for Efficient In-Core Path Tracing
Replies: 14
Views: 14799

Re: Improving Data Locality for Efficient In-Core Path Traci

Hello,
Please send me a copy of your paper Jacco.
by ziu
Fri Aug 23, 2013 6:15 pm
Forum: GPU
Topic: Speed Issues with GPU BVH rendering
Replies: 14
Views: 9549

Re: Speed Issues with GPU BVH rendering

It's hard to compare these things. Those results are from their Wavefront paper, and their code is designed to work with complex shaders. The number I was mentioning was for primary rays with very basic shading, since your shot is visualizing only traversal depth. Other things to take into consider...
by ziu
Fri Aug 23, 2013 5:42 pm
Forum: GPU
Topic: Speed Issues with GPU BVH rendering
Replies: 14
Views: 9549

Re: Speed Issues with GPU BVH rendering

I'm afraid you're much further behind atm than 5x. :) On a single 670, it's quite possible to do 300Mrays/s for a scene like that and shading like that. Sure. But I am talking about ray tracing without using bundles. Or specialized triangles. Or ray reordering. Or persistent threads. Not to mention...
by ziu
Fri Aug 23, 2013 12:59 am
Forum: GPU
Topic: Speed Issues with GPU BVH rendering
Replies: 14
Views: 9549

Re: Speed Issues with GPU BVH rendering

Yay! I figured it out! The leaf lists are lists of keys to primitive indexes! Not the direct indexes! So you need to use the key to get the correct value! It was a wonder anything rendered at all. Ah the wonders of trying to use someone else's code... :) http://i.imgur.com/clIFJRt.png?1 So now it is...
by ziu
Thu Aug 22, 2013 4:13 pm
Forum: GPU
Topic: Speed Issues with GPU BVH rendering
Replies: 14
Views: 9549

Re: Speed Issues with GPU BVH rendering

I want to have fast rebuilds so I can support animated scenes quickly. This is why I looked at the Morton code BVH GPU tree construction methods. SBVH takes seconds to construct a BVH while LBVH takes tens of milliseconds. Sure you lose somewhat on rendering speed but its a tradeoff. That paper by K...
by ziu
Tue Aug 20, 2013 9:26 pm
Forum: GPU
Topic: Speed Issues with GPU BVH rendering
Replies: 14
Views: 9549

Re: Speed Issues with GPU BVH rendering

Try Aila & Laine's traversal code. It's fast. Assuming your BVH is 'good', you should be able to get hundreds of millions of rays per second. I looked at their code. They are using a bog standard SAH BVH construction algorithm (takes whole seconds instead of tens of milliseconds like this one t...
by ziu
Tue Aug 20, 2013 9:13 pm
Forum: GPU
Topic: Speed Issues with GPU BVH rendering
Replies: 14
Views: 9549

Speed Issues with GPU BVH rendering

Hello, I am presently trying to evaluate the performance of best of class GPU BVH ray tracing algorithms for real-time rendering. Of these the work by Luebke, Pantaleoni, Karras, and Garanzha stood out. The only open source implementation I could find of these was Pantaleoni's HLBVH which is provide...

Go to advanced search