Clean up mzd_process_rows and _mzd_combine8

Issue #20 resolved
Martin Albrecht repo owner created an issue

//mzd_combine4// and //mzd_process_rows4// do essentially the same work, but the former uses SSE2 if available. They should be unified to //mzd_combineX// which can be used in both PLS decomposition, echelon forms and multiplication. Note that extensive testing must be performed to ensure there is no performance regression.

Comments (1)

  1. Log in to comment