Optimization of PCL point cloud transform functions using SSE2/AVX intrinsics.

1. Clone this repository with submodules (pass --recursive flag)

2. Configure with CMake

3. Build and run benchmarks: make benchmarks

4. Build and run tests: make tests

Columns:

• Dense: dense point clouds (thus no finity checks in transformation loop)
• Sparse: sparse point clouds with 50% of invalid points
• No copy: transform function only sets point coordinates, other fields are left uninitialized
• Copy all data: transform function copies all fields and thes sets point coordinates
• XYZ: point clouds with pcl::PointXYZ points
• XYZRGBNormal: point clouds with pcl::PointXYZRGBNormal points

Rows:

• none: no compiler flags (which means SSE2 are enabled on 64-bit machines)
• native: -march=native, which means SSE2, AVX, FMA, etc. are enabled
• sse2: -msse2
• no-sse2: -mno-sse2

For every cell two tests are run: baseline PCL transform and proposed transform. The input as a random VGA-sized point cloud. The reported number is the rutime (in microseconds) for the proposed transform and the fraction of the baseline time. Thus numbers less than 1.0 represent improved performance.