SLATEC Routines --- DSDOMN ---


*DECK DSDOMN
      SUBROUTINE DSDOMN (N, B, X, NELT, IA, JA, A, ISYM, NSAVE, ITOL,
     +   TOL, ITMAX, ITER, ERR, IERR, IUNIT, RWORK, LENW, IWORK, LENIW)
C***BEGIN PROLOGUE  DSDOMN
C***PURPOSE  Diagonally Scaled Orthomin Sparse Iterative Ax=b Solver.
C            Routine to solve a general linear system  Ax = b  using
C            the Orthomin method with diagonal scaling.
C***LIBRARY   SLATEC (SLAP)
C***CATEGORY  D2A4, D2B4
C***TYPE      DOUBLE PRECISION (SSDOMN-S, DSDOMN-D)
C***KEYWORDS  ITERATIVE PRECONDITION, NON-SYMMETRIC LINEAR SYSTEM SOLVE,
C             SLAP, SPARSE
C***AUTHOR  Greenbaum, Anne, (Courant Institute)
C           Seager, Mark K., (LLNL)
C             Lawrence Livermore National Laboratory
C             PO BOX 808, L-60
C             Livermore, CA 94550 (510) 423-3141
C             seager@llnl.gov
C***DESCRIPTION
C
C *Usage:
C     INTEGER N, NELT, IA(NELT), JA(NELT), ISYM, NSAVE, ITOL, ITMAX
C     INTEGER ITER, IERR, IUNIT, LENW, IWORK(10), LENIW
C     DOUBLE PRECISION B(N), X(N), A(NELT), TOL, ERR
C     DOUBLE PRECISION RWORK(7*N+3*N*NSAVE+NSAVE)
C
C     CALL DSDOMN(N, B, X, NELT, IA, JA, A, ISYM, NSAVE, ITOL, TOL,
C    $     ITMAX, ITER, ERR, IERR, IUNIT, RWORK, LENW, IWORK, LENIW )
C
C *Arguments:
C N      :IN       Integer.
C         Order of the Matrix.
C B      :IN       Double Precision B(N).
C         Right-hand side vector.
C X      :INOUT    Double Precision X(N).
C         On input X is your initial guess for solution vector.
C         On output X is the final approximate solution.
C NELT   :IN       Integer.
C         Number of Non-Zeros stored in A.
C IA     :IN       Integer IA(NELT).
C JA     :IN       Integer JA(NELT).
C A      :IN       Double Precision A(NELT).
C         These arrays should hold the matrix A in either the SLAP
C         Triad format or the SLAP Column format.  See "Description",
C         below.  If the SLAP Triad format is chosen, it is changed
C         internally to the SLAP Column format.
C ISYM   :IN       Integer.
C         Flag to indicate symmetric storage format.
C         If ISYM=0, all non-zero entries of the matrix are stored.
C         If ISYM=1, the matrix is symmetric, and only the upper
C         or lower triangle of the matrix is stored.
C NSAVE  :IN       Integer.
C         Number of direction vectors to save and orthogonalize against.
C ITOL   :IN       Integer.
C         Flag to indicate type of convergence criterion.
C         If ITOL=1, iteration stops when the 2-norm of the residual
C         divided by the 2-norm of the right-hand side is less than TOL.
C         If ITOL=2, iteration stops when the 2-norm of M-inv times the
C         residual divided by the 2-norm of M-inv times the right hand
C         side is less than TOL, where M-inv is the inverse of the
C         diagonal of A.
C         ITOL=11 is often useful for checking and comparing different
C         routines.  For this case, the user must supply the "exact"
C         solution or a very accurate approximation (one with an error
C         much less than TOL) through a common block,
C             COMMON /DSLBLK/ SOLN( )
C         If ITOL=11, iteration stops when the 2-norm of the difference
C         between the iterative approximation and the user-supplied
C         solution divided by the 2-norm of the user-supplied solution
C         is less than TOL.
C TOL    :INOUT    Double Precision.
C         Convergence criterion, as described above.  (Reset if IERR=4.)
C ITMAX  :IN       Integer.
C         Maximum number of iterations.
C ITER   :OUT      Integer.
C         Number of iterations required to reach convergence, or
C         ITMAX+1 if convergence criterion could not be achieved in
C         ITMAX iterations.
C ERR    :OUT      Double Precision.
C         Error estimate of error in final approximate solution, as
C         defined by ITOL.
C IERR   :OUT      Integer.
C         Return error flag.
C           IERR = 0 => All went well.
C           IERR = 1 => Insufficient space allocated for WORK or IWORK.
C           IERR = 2 => Method failed to converge in ITMAX steps.
C           IERR = 3 => Error in user input.
C                       Check input values of N, ITOL.
C           IERR = 4 => User error tolerance set too tight.
C                       Reset to 500*D1MACH(3).  Iteration proceeded.
C           IERR = 5 => Preconditioning matrix, M, is not positive
C                       definite.  (r,z) < 0.
C           IERR = 6 => Breakdown of method detected.
C                       (p,Ap) < epsilon**2.
C IUNIT  :IN       Integer.
C         Unit number on which to write the error at each iteration,
C         if this is desired for monitoring convergence.  If unit
C         number is 0, no writing will occur.
C RWORK  :WORK     Double Precision RWORK(LENW).
C         Double Precision array used for workspace.
C LENW   :IN       Integer.
C         Length of the double precision workspace, RWORK.
C         LENW >= 7*N+NSAVE*(3*N+1).
C IWORK  :WORK     Integer IWORK(LENIW).
C         Used to hold pointers into the RWORK array.
C LENIW  :IN       Integer.
C         Length of the integer workspace, IWORK.  LENIW >= 10.
C
C *Description:
C       This routine  is simply a driver  for  the DOMN routine.  It
C       calls the DSDS  routine  to set  up the  preconditioning and
C       then   calls DOMN with the   appropriate   MATVEC and MSOLVE
C       routines.
C
C       The Sparse Linear Algebra Package (SLAP) utilizes two matrix
C       data structures: 1) the  SLAP Triad  format or  2)  the SLAP
C       Column format.  The user can hand this routine either of the
C       of these data structures and SLAP  will figure out  which on
C       is being used and act accordingly.
C
C       =================== S L A P Triad format ===================
C
C       In  this   format only the  non-zeros are  stored.  They may
C       appear  in *ANY* order.   The user  supplies three arrays of
C       length NELT, where  NELT  is the number  of non-zeros in the
C       matrix:  (IA(NELT), JA(NELT),  A(NELT)).  For each  non-zero
C       the  user puts   the row  and  column index   of that matrix
C       element in the IA and JA arrays.  The  value of the non-zero
C       matrix  element is  placed in  the corresponding location of
C       the A  array.  This is  an extremely easy data  structure to
C       generate.  On  the other hand it  is  not too  efficient  on
C       vector  computers   for the  iterative  solution  of  linear
C       systems.  Hence, SLAP  changes this input  data structure to
C       the SLAP   Column  format for the  iteration (but   does not
C       change it back).
C
C       Here is an example of the  SLAP Triad   storage format for a
C       5x5 Matrix.  Recall that the entries may appear in any order.
C
C           5x5 Matrix      SLAP Triad format for 5x5 matrix on left.
C                              1  2  3  4  5  6  7  8  9 10 11
C       |11 12  0  0 15|   A: 51 12 11 33 15 53 55 22 35 44 21
C       |21 22  0  0  0|  IA:  5  1  1  3  1  5  5  2  3  4  2
C       | 0  0 33  0 35|  JA:  1  2  1  3  5  3  5  2  5  4  1
C       | 0  0  0 44  0|
C       |51  0 53  0 55|
C
C       =================== S L A P Column format ==================
C
C       In  this format   the non-zeros are    stored counting  down
C       columns (except  for the diagonal  entry, which must  appear
C       first  in each "column") and are  stored in the  double pre-
C       cision array  A. In  other  words,  for each  column  in the
C       matrix  first put  the diagonal entry in A.  Then put in the
C       other non-zero  elements going  down the column  (except the
C       diagonal)  in order.  The IA array  holds the  row index for
C       each non-zero.  The JA array  holds the offsets into the IA,
C       A  arrays  for  the  beginning  of  each  column.  That  is,
C       IA(JA(ICOL)),A(JA(ICOL)) are the first elements of the ICOL-
C       th column in IA and A, and IA(JA(ICOL+1)-1), A(JA(ICOL+1)-1)
C       are  the last elements of the ICOL-th column.   Note that we
C       always have JA(N+1)=NELT+1, where N is the number of columns
C       in the matrix  and NELT  is the number  of non-zeros  in the
C       matrix.
C
C       Here is an example of the  SLAP Column  storage format for a
C       5x5 Matrix (in the A and IA arrays '|'  denotes the end of a
C       column):
C
C           5x5 Matrix      SLAP Column format for 5x5 matrix on left.
C                              1  2  3    4  5    6  7    8    9 10 11
C       |11 12  0  0 15|   A: 11 21 51 | 22 12 | 33 53 | 44 | 55 15 35
C       |21 22  0  0  0|  IA:  1  2  5 |  2  1 |  3  5 |  4 |  5  1  3
C       | 0  0 33  0 35|  JA:  1  4  6    8  9   12
C       | 0  0  0 44  0|
C       |51  0 53  0 55|
C
C *Side Effects:
C       The SLAP Triad format (IA, JA, A)  is modified internally to
C       be the SLAP Column format.  See above.
C
C *Cautions:
C     This routine will attempt to write to the Fortran logical output
C     unit IUNIT, if IUNIT .ne. 0.  Thus, the user must make sure that
C     this logical unit is attached to a file or terminal before calling
C     this routine with a non-zero value for IUNIT.  This routine does
C     not check for the validity of a non-zero IUNIT unit number.
C
C***SEE ALSO  DOMN, DSLUOM
C***REFERENCES  (NONE)
C***ROUTINES CALLED  DCHKW, DOMN, DS2Y, DSDI, DSDS, DSMV
C***REVISION HISTORY  (YYMMDD)
C   890404  DATE WRITTEN
C   890404  Previous REVISION DATE
C   890915  Made changes requested at July 1989 CML Meeting.  (MKS)
C   890921  Removed TeX from comments.  (FNF)
C   890922  Numerous changes to prologue to make closer to SLATEC
C           standard.  (FNF)
C   890929  Numerous changes to reduce SP/DP differences.  (FNF)
C   910411  Prologue converted to Version 4.0 format.  (BAB)
C   920407  COMMON BLOCK renamed DSLBLK.  (WRB)
C   920511  Added complete declaration section.  (WRB)
C   921113  Corrected C***CATEGORY line.  (FNF)
C***END PROLOGUE  DSDOMN