Implements the paper "Understanding the Efficiency of Ray Traversal on GPUs".
Timo Aila and Samuli Laine,
Proc. High-Performance Graphics 2009
Uses NVIDIA PTX ballot instructions. This is based on:
https://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/source/browse/trunk/src/rt/kernels/kepler_dynamic_fetch.cu
Did not prove to be faster in the tests I ran compared with our simple static ray scheduler.
Patch is included for reference.