11 4 we validate the algorithm by hand applying it to sequential versions of parallel fortran programs operating over dense matrices