To solve large-scale matrix eigenvalue problems (EVPs), a two-step tridiagonalization method using the block Householder transformation (HT) is often employed. Although based on dense arithmetic requires memory storage of O(N2) and an operations O(N3), in this study, these were reduced by approximating low-rank (BLR-) matrices. A special HT for BLR-matrices it are proposed to EVP with real symm...
