ray tracer choosing tools

Sure I could give you the Demo.

You can get my email ID from my profile.

Cheers

Akshay

Does anybody have kd-tree generation code they could share? Or some examples to be found on the net, which I can use to build on? I am having trouble generating kd trees from my geometry.

I should say, I have a mixture of big & small triangles, and quite some of them are axis-aligned (say, 25-50%).

For me the kd-tree with ropes still looks extremely useful, partly because I will probably have a lot of rays, that will terminate quite quickly after the source point, and also because it has already been implemented in CUDA according to the paper, so it should be feasible ;)

Hi,

has anybody implemented the paper “Realtime Raytracing on GPU with BVH-based Packet Traversal” or “Stackless KD-Tree Traversal for High Performance GPU Ray Tracing” from Günther, Popov, Seidel and Slusallek?

They always write something of “Number of processor” or “Processor ID”… I understand it like I have to take the actual number of the running multiprocessor…Am I right or is mean something else?

And how could I take the ID of the actual running processor?

Thanks and greetings,

thopil :thumbup:

I will be trying to implement the stackless traversal algorithm. So maybe I can tell you something later this week (he said optimistically ;))

And for what it’s worth: after looking at the yafray kdtree generation code, I now have a working kdtree-generator in matlab. Now I just need to optimize my rope-generation, and then implement the traversal (first in matlab, then in CUDA)

I will probably also be trying to implement a BVH based algorithm to compare, since the kd-trees I generate take up to 3 times as many planes to store than the number of planes in the scene, although that will probably improve when converting everything into triangles, like I need to do anyway for calculation on GPU.

hey buddy,
Real-Time KD-Tree Construction on Graphics Hardware

Kun Zhou, Qiming Hou, Rui Wang, Baining Guo.

ACM TOG (conditionally accepted, pending major revisions),

http://research.microsoft.com/users/kunzho…-TR-2008-52.pdf