dppsvx (l) - Linux Manuals

dppsvx: uses the Cholesky factorization A = U**T*U or A = L*L**T to compute the solution to a real system of linear equations A * X = B,

NAME

DPPSVX - uses the Cholesky factorization A = U**T*U or A = L*L**T to compute the solution to a real system of linear equations A * X = B,

SYNOPSIS

SUBROUTINE DPPSVX(

FACT, UPLO, N, NRHS, AP, AFP, EQUED, S, B, LDB, X, LDX, RCOND, FERR, BERR, WORK, IWORK, INFO )

    

CHARACTER EQUED, FACT, UPLO

    

INTEGER INFO, LDB, LDX, N, NRHS

    

DOUBLE PRECISION RCOND

    

INTEGER IWORK( * )

    

DOUBLE PRECISION AFP( * ), AP( * ), B( LDB, * ), BERR( * ), FERR( * ), S( * ), WORK( * ), X( LDX, * )

PURPOSE

DPPSVX uses the Cholesky factorization A = U**T*U or A = L*L**T to compute the solution to a real system of linear equations
A * X = B, where A is an N-by-N symmetric positive definite matrix stored in packed format and X and B are N-by-NRHS matrices.
Error bounds on the solution and a condition estimate are also provided.

DESCRIPTION

The following steps are performed:
1. If FACT = aqEaq, real scaling factors are computed to equilibrate
the system:

diag(S) * A * diag(S) * inv(diag(S)) * X = diag(S) * B
Whether or not the system will be equilibrated depends on the
scaling of the matrix A, but if equilibration is used, A is
overwritten by diag(S)*A*diag(S) and B by diag(S)*B.
2. If FACT = aqNaq or aqEaq, the Cholesky decomposition is used to
factor the matrix A (after equilibration if FACT = aqEaq) as
A = U**T* U,  if UPLO = aqUaq, or

A = L * L**T,  if UPLO = aqLaq,

where U is an upper triangular matrix and L is a lower triangular
matrix.
3. If the leading i-by-i principal minor is not positive definite,
then the routine returns with INFO = i. Otherwise, the factored
form of A is used to estimate the condition number of the matrix
A.  If the reciprocal of the condition number is less than machine
precision, INFO = N+1 is returned as a warning, but the routine
still goes on to solve for X and compute error bounds as
described below.
4. The system of equations is solved for X using the factored form
of A.
5. Iterative refinement is applied to improve the computed solution
matrix and calculate error bounds and backward error estimates
for it.
6. If equilibration was used, the matrix X is premultiplied by
diag(S) so that it solves the original system before

equilibration.

ARGUMENTS

FACT (input) CHARACTER*1

Specifies whether or not the factored form of the matrix A is supplied on entry, and if not, whether the matrix A should be equilibrated before it is factored. = aqFaq: On entry, AFP contains the factored form of A. If EQUED = aqYaq, the matrix A has been equilibrated with scaling factors given by S. AP and AFP will not be modified. = aqNaq: The matrix A will be copied to AFP and factored.
= aqEaq: The matrix A will be equilibrated if necessary, then copied to AFP and factored.

UPLO (input) CHARACTER*1

= aqUaq: Upper triangle of A is stored;
= aqLaq: Lower triangle of A is stored.

N (input) INTEGER

The number of linear equations, i.e., the order of the matrix A. N >= 0.

NRHS (input) INTEGER

The number of right hand sides, i.e., the number of columns of the matrices B and X. NRHS >= 0.

AP (input/output) DOUBLE PRECISION array, dimension (N*(N+1)/2)

On entry, the upper or lower triangle of the symmetric matrix A, packed columnwise in a linear array, except if FACT = aqFaq and EQUED = aqYaq, then A must contain the equilibrated matrix diag(S)*A*diag(S). The j-th column of A is stored in the array AP as follows: if UPLO = aqUaq, AP(i + (j-1)*j/2) = A(i,j) for 1<=i<=j; if UPLO = aqLaq, AP(i + (j-1)*(2n-j)/2) = A(i,j) for j<=i<=n. See below for further details. A is not modified if FACT = aqFaq or aqNaq, or if FACT = aqEaq and EQUED = aqNaq on exit. On exit, if FACT = aqEaq and EQUED = aqYaq, A is overwritten by diag(S)*A*diag(S).

AFP (input or output) DOUBLE PRECISION array, dimension

(N*(N+1)/2) If FACT = aqFaq, then AFP is an input argument and on entry contains the triangular factor U or L from the Cholesky factorization A = Uaq*U or A = L*Laq, in the same storage format as A. If EQUED .ne. aqNaq, then AFP is the factored form of the equilibrated matrix A. If FACT = aqNaq, then AFP is an output argument and on exit returns the triangular factor U or L from the Cholesky factorization A = Uaq*U or A = L*Laq of the original matrix A. If FACT = aqEaq, then AFP is an output argument and on exit returns the triangular factor U or L from the Cholesky factorization A = Uaq*U or A = L*Laq of the equilibrated matrix A (see the description of AP for the form of the equilibrated matrix).

EQUED (input or output) CHARACTER*1

Specifies the form of equilibration that was done. = aqNaq: No equilibration (always true if FACT = aqNaq).
= aqYaq: Equilibration was done, i.e., A has been replaced by diag(S) * A * diag(S). EQUED is an input argument if FACT = aqFaq; otherwise, it is an output argument.

S (input or output) DOUBLE PRECISION array, dimension (N)

The scale factors for A; not accessed if EQUED = aqNaq. S is an input argument if FACT = aqFaq; otherwise, S is an output argument. If FACT = aqFaq and EQUED = aqYaq, each element of S must be positive.

B (input/output) DOUBLE PRECISION array, dimension (LDB,NRHS)

On entry, the N-by-NRHS right hand side matrix B. On exit, if EQUED = aqNaq, B is not modified; if EQUED = aqYaq, B is overwritten by diag(S) * B.

LDB (input) INTEGER

The leading dimension of the array B. LDB >= max(1,N).

X (output) DOUBLE PRECISION array, dimension (LDX,NRHS)

If INFO = 0 or INFO = N+1, the N-by-NRHS solution matrix X to the original system of equations. Note that if EQUED = aqYaq, A and B are modified on exit, and the solution to the equilibrated system is inv(diag(S))*X.

LDX (input) INTEGER

The leading dimension of the array X. LDX >= max(1,N).

RCOND (output) DOUBLE PRECISION

The estimate of the reciprocal condition number of the matrix A after equilibration (if done). If RCOND is less than the machine precision (in particular, if RCOND = 0), the matrix is singular to working precision. This condition is indicated by a return code of INFO > 0.

FERR (output) DOUBLE PRECISION array, dimension (NRHS)

The estimated forward error bound for each solution vector X(j) (the j-th column of the solution matrix X). If XTRUE is the true solution corresponding to X(j), FERR(j) is an estimated upper bound for the magnitude of the largest element in (X(j) - XTRUE) divided by the magnitude of the largest element in X(j). The estimate is as reliable as the estimate for RCOND, and is almost always a slight overestimate of the true error.

BERR (output) DOUBLE PRECISION array, dimension (NRHS)

The componentwise relative backward error of each solution vector X(j) (i.e., the smallest relative change in any element of A or B that makes X(j) an exact solution).

WORK (workspace) DOUBLE PRECISION array, dimension (3*N)

IWORK (workspace) INTEGER array, dimension (N)

INFO (output) INTEGER

= 0: successful exit
< 0: if INFO = -i, the i-th argument had an illegal value
> 0: if INFO = i, and i is
<= N: the leading minor of order i of A is not positive definite, so the factorization could not be completed, and the solution has not been computed. RCOND = 0 is returned. = N+1: U is nonsingular, but RCOND is less than machine precision, meaning that the matrix is singular to working precision. Nevertheless, the solution and error bounds are computed because there are a number of situations where the computed solution can be more accurate than the value of RCOND would suggest.

FURTHER DETAILS

The packed storage scheme is illustrated by the following example when N = 4, UPLO = aqUaq:
Two-dimensional storage of the symmetric matrix A:

a11 a12 a13 a14

 a22 a23 a24

     a33 a34     (aij = conjg(aji))

         a44
Packed storage of the upper triangle of A:
AP = [ a11, a12, a22, a13, a23, a33, a14, a24, a34, a44 ]