cuda::experimental::places::owners_to_block_runs

inline ::std::vector<block_run> cuda::experimental::places::owners_to_block_runs(

const ::std::vector<pos4> &owners

)

Merge a one-owner-per-block vector into maximal block runs.