I've tried to keep the structure of this pretty flat. The folders here are:
GSseq - where I initially played around with GS in c. WARNING: code here may be incorrect.
GSseqcpp - where I moved GSseq to c++. WARNING: code here may be incorrect
GScuda - where I worked implementing GS in CUDA. WARNING: code here may be incorrect.
small_matching - GS and PII implentations both sequentially and in CUDA for problems with n < 200. WARNING code here may be incorrect.
matching - sort of the final destination for my cuda expirimentation and sequential GS stuff. The GS code here is probably correct and is where I got my data from.
{sanity_check,analysis,NPGS} - clones of matching with slight modification for different purposes.
expr - some idea I was playing around with that had something to do with collisions. You can probably ignore thing.
graph-gen - some code to make graphs in DOT format of how divorces and PII traverse the space of all possible matchings.
rvvpii - started as an extension of graphs, but turned into my implementation of CPII and PII that I got my data from. 


Pretty much everything was compiled in CentOS and/or ubuntu with gcc and nvcc.
You really should just need to run the make files to compile things.
All of this code was just me playing around with ideas so most of it is pretty unpolished,
although I will stand by the reliability of the code in the rvvpii and matching folders that 
I got my data from.

You can contact me with quesions at: caleb.clark@my.wheaton.edu