|
180. NESBET (Version A+B): Lowest Eigenvalue and
Eigenvector of Large and Very Large Matrices
by M. Raimondi and G. F. Tantardini, Physical Chemistry Institute, University of Milan, Milan, Italy Subroutine NESBET (Version A) computes the lowest eigenvalue and eigenvector of a very large, real symmetric matrix H when the corresponding orthogonality matrix S in the secular (H-E*S) *C=0 is real, symmetric and positive definite. The two matrices S and H are stored in auxiliary memory (disk or magnetic tape). Version B is used for large matrices and is completely core contained as opposted to Version A which uses disk or tape. FORTRAN IV (UNIVAC 1108) Lines of Code: 300 Recommended Citation:M. Raimondi and G. F. Tantardini, QCPE 11, 180 (1971). |