MATRIX commands


PROGRAM NAME

minvert - Compute Inverse Matrix

DESCRIPTION

minvert computes the inverse of a matrix. If the -diag flag is selected, then only the diagonal elements are inverted (this will work for any size of matrix). If the -diag flag is not selected, then the matrix must be square, and the inverse will be computed by backsolving an LU decomposition.

minvert uses the LAPACK routines DGETRF, DGETRI, ZGERF, and ZGTRI to actually perform the decomposition and compute the inverse.

The output data type will be KDCOMPLEX if the input matrix is either COMPLEX or KDCOMPLEX. Otherwise the output will be KDOUBLE.

After execution of the appropriate LAPACK routines, a diagnostic may be printed out indicating that the input matrix is singular and at what point this was detected. This will be printed only if KHOROS_NOTIFY is set to STANDARD or more verbosity. This message indicates a fatal error.

REQUIRED ARGUMENTS

-i
type: infile
desc: Input matrix
-o
type: outfile
desc: Output matrix

OPTIONAL ARGUMENTS

-diag
type: flag
desc: Invert diagonal only

EXAMPLES

SEE ALSO

RESTRICTIONS

minvert does not check for a divide by zero condition when computing the inverse with the -diag flag selected.

REFERENCES

LAPACK Users' Guide, E. Anderson et. al., SIAM 1992. ISBN 0-89871-294-7

COPYRIGHT

Copyright (C) 1993 - 1997, Khoral Research, Inc. ("KRI") All rights reserved.