I have a triangle mesh. For each vertex of that mesh i want to determine, wether it is visible from the a camera position or not, i.e. wether it the vertex is inside the viewing frustum, showing its front-face and is not occluded by parts of the mesh (triangle-) surface. With a GPU this is trivial and fast.
However, now i am interested in how fast this problem could be solved on a cpu by applying whatever kind of algorithm. (to get an idea of the CPU-GPU gap for this problem and underline the superior speed of the GPU (if there is) )
Could anyone point me to
- current literature where such or very similar comparisons have been done
- to state-of-the-art implementations of CPU-based algorithms in that area that are or could easily be used to create benchmarks for this?