I have been reading Timo's paper: understanding-the-efficiency-of-ray-traversal-on-gpus
In which it mentioned that speculative traversal will help the performance.
I read that paragraph several times and still don't know what's doing.
Is it keeping all the rays in wavefront? or simply keeping tracing rays until all reached leafs?
If keep tracing, what to tracing when leaf is already reached.
I think that post has omitted so many details if you want to implement a fast GPU tracer from ground up.
Such as seperating ray generation and ray traversal. And tracing rays in batches (what the best batch size?)
How to sort rays, and so on.
Any one can help more?
Practical and theoretical implementation discussion.
1 post • Page 1 of 1
Who is online
Users browsing this forum: No registered users and 7 guests