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.