Hi I am Abhilash from India.I recently applied a proposal for Cycle BVH traversal optimisation based on SAH and faster object splitting based on orientation (cones of normals) and light source intensity.I have a C++ prototype for allowing stochastic distribution of the light intensity as a probability factor for heavy lights, and splitting the object nodes .However there is a significant tradeoff between the execution time (traversal BVH) and time owing to BVH splitting . I have a draft proposal ready and the original proposal is submitted:
Can a possible feedback be provided as to the content of the proposal based on SBVH traversal and sampling.I recall the deadline has passed but I would be greatful if Brecht van Lommel ,Sergey Sharybin gave a feedback .Plus I have also drafter another proposal based on Embree STBVH optimisation over Cycle BVH based on MBSAH for faster motion blur.
The link for second proposal is here:
Thank s for any feedback