1       SUBROUTINE DGBSV( N, KL, KU, NRHS, AB, LDAB, IPIV, B, LDB, INFO )
  2 *
  3 *  -- LAPACK driver routine (version 3.2) --
  4 *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  5 *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  6 *     November 2006
  7 *
  8 *     .. Scalar Arguments ..
  9       INTEGER            INFO, KL, KU, LDAB, LDB, N, NRHS
 10 *     ..
 11 *     .. Array Arguments ..
 12       INTEGER            IPIV( * )
 13       DOUBLE PRECISION   AB( LDAB, * ), B( LDB, * )
 14 *     ..
 15 *
 16 *  Purpose
 17 *  =======
 18 *
 19 *  DGBSV computes the solution to a real system of linear equations
 20 *  A * X = B, where A is a band matrix of order N with KL subdiagonals
 21 *  and KU superdiagonals, and X and B are N-by-NRHS matrices.
 22 *
 23 *  The LU decomposition with partial pivoting and row interchanges is
 24 *  used to factor A as A = L * U, where L is a product of permutation
 25 *  and unit lower triangular matrices with KL subdiagonals, and U is
 26 *  upper triangular with KL+KU superdiagonals.  The factored form of A
 27 *  is then used to solve the system of equations A * X = B.
 28 *
 29 *  Arguments
 30 *  =========
 31 *
 32 *  N       (input) INTEGER
 33 *          The number of linear equations, i.e., the order of the
 34 *          matrix A.  N >= 0.
 35 *
 36 *  KL      (input) INTEGER
 37 *          The number of subdiagonals within the band of A.  KL >= 0.
 38 *
 39 *  KU      (input) INTEGER
 40 *          The number of superdiagonals within the band of A.  KU >= 0.
 41 *
 42 *  NRHS    (input) INTEGER
 43 *          The number of right hand sides, i.e., the number of columns
 44 *          of the matrix B.  NRHS >= 0.
 45 *
 46 *  AB      (input/output) DOUBLE PRECISION array, dimension (LDAB,N)
 47 *          On entry, the matrix A in band storage, in rows KL+1 to
 48 *          2*KL+KU+1; rows 1 to KL of the array need not be set.
 49 *          The j-th column of A is stored in the j-th column of the
 50 *          array AB as follows:
 51 *          AB(KL+KU+1+i-j,j) = A(i,j) for max(1,j-KU)<=i<=min(N,j+KL)
 52 *          On exit, details of the factorization: U is stored as an
 53 *          upper triangular band matrix with KL+KU superdiagonals in
 54 *          rows 1 to KL+KU+1, and the multipliers used during the
 55 *          factorization are stored in rows KL+KU+2 to 2*KL+KU+1.
 56 *          See below for further details.
 57 *
 58 *  LDAB    (input) INTEGER
 59 *          The leading dimension of the array AB.  LDAB >= 2*KL+KU+1.
 60 *
 61 *  IPIV    (output) INTEGER array, dimension (N)
 62 *          The pivot indices that define the permutation matrix P;
 63 *          row i of the matrix was interchanged with row IPIV(i).
 64 *
 65 *  B       (input/output) DOUBLE PRECISION array, dimension (LDB,NRHS)
 66 *          On entry, the N-by-NRHS right hand side matrix B.
 67 *          On exit, if INFO = 0, the N-by-NRHS solution matrix X.
 68 *
 69 *  LDB     (input) INTEGER
 70 *          The leading dimension of the array B.  LDB >= max(1,N).
 71 *
 72 *  INFO    (output) INTEGER
 73 *          = 0:  successful exit
 74 *          < 0:  if INFO = -i, the i-th argument had an illegal value
 75 *          > 0:  if INFO = i, U(i,i) is exactly zero.  The factorization
 76 *                has been completed, but the factor U is exactly
 77 *                singular, and the solution has not been computed.
 78 *
 79 *  Further Details
 80 *  ===============
 81 *
 82 *  The band storage scheme is illustrated by the following example, when
 83 *  M = N = 6, KL = 2, KU = 1:
 84 *
 85 *  On entry:                       On exit:
 86 *
 87 *      *    *    *    +    +    +       *    *    *   u14  u25  u36
 88 *      *    *    +    +    +    +       *    *   u13  u24  u35  u46
 89 *      *   a12  a23  a34  a45  a56      *   u12  u23  u34  u45  u56
 90 *     a11  a22  a33  a44  a55  a66     u11  u22  u33  u44  u55  u66
 91 *     a21  a32  a43  a54  a65   *      m21  m32  m43  m54  m65   *
 92 *     a31  a42  a53  a64   *    *      m31  m42  m53  m64   *    *
 93 *
 94 *  Array elements marked * are not used by the routine; elements marked
 95 *  + need not be set on entry, but are required by the routine to store
 96 *  elements of U because of fill-in resulting from the row interchanges.
 97 *
 98 *  =====================================================================
 99 *
100 *     .. External Subroutines ..
101       EXTERNAL           DGBTRF, DGBTRS, XERBLA
102 *     ..
103 *     .. Intrinsic Functions ..
104       INTRINSIC          MAX
105 *     ..
106 *     .. Executable Statements ..
107 *
108 *     Test the input parameters.
109 *
110       INFO = 0
111       IF( N.LT.0 ) THEN
112          INFO = -1
113       ELSE IF( KL.LT.0 ) THEN
114          INFO = -2
115       ELSE IF( KU.LT.0 ) THEN
116          INFO = -3
117       ELSE IF( NRHS.LT.0 ) THEN
118          INFO = -4
119       ELSE IF( LDAB.LT.2*KL+KU+1 ) THEN
120          INFO = -6
121       ELSE IF( LDB.LT.MAX( N, 1 ) ) THEN
122          INFO = -9
123       END IF
124       IF( INFO.NE.0 ) THEN
125          CALL XERBLA( 'DGBSV '-INFO )
126          RETURN
127       END IF
128 *
129 *     Compute the LU factorization of the band matrix A.
130 *
131       CALL DGBTRF( N, N, KL, KU, AB, LDAB, IPIV, INFO )
132       IF( INFO.EQ.0 ) THEN
133 *
134 *        Solve the system A*X = B, overwriting B with X.
135 *
136          CALL DGBTRS( 'No transpose', N, KL, KU, NRHS, AB, LDAB, IPIV,
137      $                B, LDB, INFO )
138       END IF
139       RETURN
140 *
141 *     End of DGBSV
142 *
143       END