Elsevier Science Home
Computer Physics Communications Program Library
Full text online from Science Direct
Programs in Physics & Physical Chemistry
CPC Home

[Licence| Download | New Version Template] abvi_v1_0.gz(9 Kbytes)
Manuscript Title: HYMNIABLOCK: eigenvalue solver for blocked matrices.
Authors: R. Gruber
Program title: HYMBLO
Catalogue identifier: ABVI_v1_0
Distribution format: gz
Journal reference: Comput. Phys. Commun. 20(1980)421
Programming language: Fortran.
Computer: CDC CYBER 7326.
Operating system: NOS BE.
RAM: 24K words
Word size: 60
Peripherals: disc.
Keywords: General purpose, Numerical mathematics, Eigenvalue problem, Inverse vector iteration, Matrix, Symmetric matrix.
Classification: 4.8.

Nature of problem:
Multidimensional stability problems lead to huge eigenvalue problems of the type Ax = lambdaBx, which are only solvable when the matrix structures are not touched. In this package, A and B have block diagonal form.

Solution method:
The inverse vector iteration is used which converges towards the lowest eigenvalue in absolute value. By means of a shift of the spectrum it is possible to obtain any one of the modes. Decomposing A first makes the iteration part much more efficient. The NEG facility gives the number of negative eigenvalues of the shifted problem and thus enables us to know which eigenvalue has been calculated.

Both matrices, A and B, have to be real symmetric and B must be positive definite. Both must have the same block diagonal structure.

Unusual features:
The test program HYMBLO reads input data using NAMELIST which is not standard FORTRAN.