The following code:
https://github.com/InverseLight/ValoMC

This code is parallelized today using OpenMP, with a Matlab wrapper.

Easy installation instructions here:
https://inverselight.github.io/ValoMC/
3D example here:
https://inverselight.github.io/ValoMC/voxeltest.html

The job:
I'm looking to take this code (Only the 3D part, so *only* MC3D.cpp and MC3D.hpp) and parallelize it using GPU (Retain the Matlab wrapper, so this needs to be CUDA\Mex).

The deliverable: An updated C++ code which compiles as MEX. The code will use GPU for the computation and will run faster on the GPU than on the CPU. I'd like to see at least x2 speedup.

The instructions for compilation are in the Github page linked above.