PCL-normal-estimation

Parallel Programing course final project: Parallelization of normal estimation in point clouds with and without PCL.

Comparisons of serial, OpenMP and CUDA implementations can be found below and speed comparisons can be found in the report.pdf:

To compile code:

1. Place code folder (it should include both the code and CMakeLists.txt) under the scratch folder in your parallel programming cluster.
2. cd to code folder (e.g. cd Normal_Estimation_CUDA)
3. mkdir build
4. cd build
5. cmake ..
6. make
7. Run executable to see its usage