Paul sent Cpp code in an email March 2019. Can we, and should we, incorporate this into HIP sampling?