An efficient and stable technique to remove the limitation in choosing a shift in the subspace iteration method with shifting is presented. A major difficulty of the subspace iteration method with shifting is that, because of the singularity problem, a shift close to an eigenvalue cannot be used, resulting in slower convergence. This study solves the above singularity problem using side conditions without sacrifice of convergence. The method is always nonsingular even if a shift is an eigenvalue itself This is one of the significant characteristics of the proposed method. The nonsingularity is proved analytically. The convergence of the proposed method is at least equal to that of the subspace iteration method with shifting, and the operation counts of above two methods are almost the same for large structures. To show the effectiveness of the proposed method, two numerical examples are considered. (C) 1999 Elsevier Science Ltd. Ail rights reserved.