f07jvf computes error bounds and refines the solution to a complex system of linear equations , where is an Hermitian positive definite tridiagonal matrix and and are matrices, using the modified Cholesky factorization returned by f07jrf and an initial solution returned by f07jsf. Iterative refinement is used to reduce the backward error as much as possible.
The routine may be called by the names f07jvf, nagf_lapacklin_zptrfs or its LAPACK name zptrfs.
3Description
f07jvf should normally be preceded by calls to f07jrfandf07jsf. f07jrf computes a modified Cholesky factorization of the matrix as
where is a unit lower bidiagonal matrix and is a diagonal matrix, with positive diagonal elements. f07jsf then utilizes the factorization to compute a solution, , to the required equations. Letting denote a column of , f07jvf computes a component-wise backward error, , the smallest relative perturbation in each element of and such that is the exact solution of a perturbed system
The routine also estimates a bound for the component-wise forward error in the computed solution defined by , where is the corresponding column of the exact solution, .
Note that the modified Cholesky factorization of can also be expressed as
where is unit upper bidiagonal.
4References
Anderson E, Bai Z, Bischof C, Blackford S, Demmel J, Dongarra J J, Du Croz J J, Greenbaum A, Hammarling S, McKenney A and Sorensen D (1999) LAPACK Users' Guide (3rd Edition) SIAM, Philadelphia https://www.netlib.org/lapack/lug
5Arguments
1: – Character(1)Input
On entry: specifies the form of the factorization as follows:
.
.
Constraint:
or .
2: – IntegerInput
On entry: , the order of the matrix .
Constraint:
.
3: – IntegerInput
On entry: , the number of right-hand sides, i.e., the number of columns of the matrix .
Constraint:
.
4: – Real (Kind=nag_wp) arrayInput
Note: the dimension of the array d
must be at least
.
On entry: must contain the diagonal elements of the matrix of .
5: – Complex (Kind=nag_wp) arrayInput
Note: the dimension of the array e
must be at least
.
On entry: if , e must contain the superdiagonal elements of the matrix .
If , e must contain the subdiagonal elements of the matrix .
6: – Real (Kind=nag_wp) arrayInput
Note: the dimension of the array df
must be at least
.
On entry: must contain the diagonal elements of the diagonal matrix from the factorization of .
7: – Complex (Kind=nag_wp) arrayInput
Note: the dimension of the array ef
must be at least
.
On entry: if , ef must contain the superdiagonal elements of the unit upper bidiagonal matrix from the factorization of .
If , ef must contain the subdiagonal elements of the unit lower bidiagonal matrix from the factorization of .
8: – Complex (Kind=nag_wp) arrayInput
Note: the second dimension of the array b
must be at least
.
On entry: the matrix of right-hand sides .
9: – IntegerInput
On entry: the first dimension of the array b as declared in the (sub)program from which f07jvf is called.
Constraint:
.
10: – Complex (Kind=nag_wp) arrayInput/Output
Note: the second dimension of the array x
must be at least
.
On entry: the initial solution matrix .
On exit: the refined solution matrix .
11: – IntegerInput
On entry: the first dimension of the array x as declared in the (sub)program from which f07jvf is called.
Constraint:
.
12: – Real (Kind=nag_wp) arrayOutput
On exit: estimate of the forward error bound for each computed solution vector, such that , where is the th column of the computed solution returned in the array x and is the corresponding column of the exact solution . The estimate is almost always a slight overestimate of the true error.
13: – Real (Kind=nag_wp) arrayOutput
On exit: estimate of the component-wise relative backward error of each computed solution vector (i.e., the smallest relative change in any element of or that makes an exact solution).
14: – Complex (Kind=nag_wp) arrayWorkspace
15: – Real (Kind=nag_wp) arrayWorkspace
16: – IntegerOutput
On exit: unless the routine detects an error (see Section 6).
6Error Indicators and Warnings
If , argument had an illegal value. An explanatory message is output, and execution of the program is terminated.
7Accuracy
The computed solution for a single right-hand side, , satisfies an equation of the form
where
and is the machine precision. An approximate error bound for the computed solution is given by
where , the condition number of with respect to the solution of the linear equations. See Section 4.4 of Anderson et al. (1999) for further details.
Routine f07juf can be used to compute the condition number of .
8Parallelism and Performance
Background information to multithreading can be found in the Multithreading documentation.
f07jvf is threaded by NAG for parallel execution in multithreaded implementations of the NAG Library.
f07jvf makes calls to BLAS and/or LAPACK routines, which may be threaded within the vendor library used by this implementation. Consult the documentation for the vendor library for further information.
Please consult the X06 Chapter Introduction for information on how to control and interrogate the OpenMP environment used within this routine. Please also consult the Users' Note for your implementation for any additional implementation-specific information.
9Further Comments
The total number of floating-point operations required to solve the equations is proportional to . At most five steps of iterative refinement are performed, but usually only one or two steps are required.