Is your feature request related to a problem? Please describe.

Benchmarks will allow developers to test their change and iterate faster simply by recompiling (and saving the binaries) and running the benchmark tests.

Context

PCL has no bench-marking, which allows some regressions to slip through.

Describe the solution you'd like

• A separate folder like tests: benchmarks
• Using a tool like google-benchmark or similar
• Bench-marking different algorithms, etc. Some examples for inspiration:
  • https://github.com/facontidavide/PCL_VoxelGrid_enhanced
  • https://github.com/taketwo/pcl-transforms