1SPTRFS(1) LAPACK routine (version 3.1) SPTRFS(1)
2
3
4
6 SPTRFS - the computed solution to a system of linear equations when the
7 coefficient matrix is symmetric positive definite and tridiagonal, and
8 provides error bounds and backward error estimates for the solution
9
11 SUBROUTINE SPTRFS( N, NRHS, D, E, DF, EF, B, LDB, X, LDX, FERR, BERR,
12 WORK, INFO )
13
14 INTEGER INFO, LDB, LDX, N, NRHS
15
16 REAL B( LDB, * ), BERR( * ), D( * ), DF( * ), E( * ), EF(
17 * ), FERR( * ), WORK( * ), X( LDX, * )
18
20 SPTRFS improves the computed solution to a system of linear equations
21 when the coefficient matrix is symmetric positive definite and tridiag‐
22 onal, and provides error bounds and backward error estimates for the
23 solution.
24
25
27 N (input) INTEGER
28 The order of the matrix A. N >= 0.
29
30 NRHS (input) INTEGER
31 The number of right hand sides, i.e., the number of columns of
32 the matrix B. NRHS >= 0.
33
34 D (input) REAL array, dimension (N)
35 The n diagonal elements of the tridiagonal matrix A.
36
37 E (input) REAL array, dimension (N-1)
38 The (n-1) subdiagonal elements of the tridiagonal matrix A.
39
40 DF (input) REAL array, dimension (N)
41 The n diagonal elements of the diagonal matrix D from the fac‐
42 torization computed by SPTTRF.
43
44 EF (input) REAL array, dimension (N-1)
45 The (n-1) subdiagonal elements of the unit bidiagonal factor L
46 from the factorization computed by SPTTRF.
47
48 B (input) REAL array, dimension (LDB,NRHS)
49 The right hand side matrix B.
50
51 LDB (input) INTEGER
52 The leading dimension of the array B. LDB >= max(1,N).
53
54 X (input/output) REAL array, dimension (LDX,NRHS)
55 On entry, the solution matrix X, as computed by SPTTRS. On
56 exit, the improved solution matrix X.
57
58 LDX (input) INTEGER
59 The leading dimension of the array X. LDX >= max(1,N).
60
61 FERR (output) REAL array, dimension (NRHS)
62 The forward error bound for each solution vector X(j) (the j-th
63 column of the solution matrix X). If XTRUE is the true solu‐
64 tion corresponding to X(j), FERR(j) is an estimated upper bound
65 for the magnitude of the largest element in (X(j) - XTRUE)
66 divided by the magnitude of the largest element in X(j).
67
68 BERR (output) REAL array, dimension (NRHS)
69 The componentwise relative backward error of each solution vec‐
70 tor X(j) (i.e., the smallest relative change in any element of
71 A or B that makes X(j) an exact solution).
72
73 WORK (workspace) REAL array, dimension (2*N)
74
75 INFO (output) INTEGER
76 = 0: successful exit
77 < 0: if INFO = -i, the i-th argument had an illegal value
78
80 ITMAX is the maximum number of steps of iterative refinement.
81
82
83
84 LAPACK routine (version 3.1) November 2006 SPTRFS(1)