File:  [local] / rpl / lapack / lapack / dpprfs.f
Revision 1.10: download - view: text, annotated - select for diffs - revision graph
Mon Nov 21 22:19:38 2011 UTC (12 years, 5 months ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_1_8, rpl-4_1_7, rpl-4_1_6, rpl-4_1_5, rpl-4_1_4, HEAD
Cohérence

    1: *> \brief \b DPPRFS
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at 
    6: *            http://www.netlib.org/lapack/explore-html/ 
    7: *
    8: *> \htmlonly
    9: *> Download DPPRFS + dependencies 
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dpprfs.f"> 
   11: *> [TGZ]</a> 
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dpprfs.f"> 
   13: *> [ZIP]</a> 
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dpprfs.f"> 
   15: *> [TXT]</a>
   16: *> \endhtmlonly 
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE DPPRFS( UPLO, N, NRHS, AP, AFP, B, LDB, X, LDX, FERR,
   22: *                          BERR, WORK, IWORK, INFO )
   23:    24: *       .. Scalar Arguments ..
   25: *       CHARACTER          UPLO
   26: *       INTEGER            INFO, LDB, LDX, N, NRHS
   27: *       ..
   28: *       .. Array Arguments ..
   29: *       INTEGER            IWORK( * )
   30: *       DOUBLE PRECISION   AFP( * ), AP( * ), B( LDB, * ), BERR( * ),
   31: *      $                   FERR( * ), WORK( * ), X( LDX, * )
   32: *       ..
   33: *  
   34: *
   35: *> \par Purpose:
   36: *  =============
   37: *>
   38: *> \verbatim
   39: *>
   40: *> DPPRFS improves the computed solution to a system of linear
   41: *> equations when the coefficient matrix is symmetric positive definite
   42: *> and packed, and provides error bounds and backward error estimates
   43: *> for the solution.
   44: *> \endverbatim
   45: *
   46: *  Arguments:
   47: *  ==========
   48: *
   49: *> \param[in] UPLO
   50: *> \verbatim
   51: *>          UPLO is CHARACTER*1
   52: *>          = 'U':  Upper triangle of A is stored;
   53: *>          = 'L':  Lower triangle of A is stored.
   54: *> \endverbatim
   55: *>
   56: *> \param[in] N
   57: *> \verbatim
   58: *>          N is INTEGER
   59: *>          The order of the matrix A.  N >= 0.
   60: *> \endverbatim
   61: *>
   62: *> \param[in] NRHS
   63: *> \verbatim
   64: *>          NRHS is INTEGER
   65: *>          The number of right hand sides, i.e., the number of columns
   66: *>          of the matrices B and X.  NRHS >= 0.
   67: *> \endverbatim
   68: *>
   69: *> \param[in] AP
   70: *> \verbatim
   71: *>          AP is DOUBLE PRECISION array, dimension (N*(N+1)/2)
   72: *>          The upper or lower triangle of the symmetric matrix A, packed
   73: *>          columnwise in a linear array.  The j-th column of A is stored
   74: *>          in the array AP as follows:
   75: *>          if UPLO = 'U', AP(i + (j-1)*j/2) = A(i,j) for 1<=i<=j;
   76: *>          if UPLO = 'L', AP(i + (j-1)*(2n-j)/2) = A(i,j) for j<=i<=n.
   77: *> \endverbatim
   78: *>
   79: *> \param[in] AFP
   80: *> \verbatim
   81: *>          AFP is DOUBLE PRECISION array, dimension (N*(N+1)/2)
   82: *>          The triangular factor U or L from the Cholesky factorization
   83: *>          A = U**T*U or A = L*L**T, as computed by DPPTRF/ZPPTRF,
   84: *>          packed columnwise in a linear array in the same format as A
   85: *>          (see AP).
   86: *> \endverbatim
   87: *>
   88: *> \param[in] B
   89: *> \verbatim
   90: *>          B is DOUBLE PRECISION array, dimension (LDB,NRHS)
   91: *>          The right hand side matrix B.
   92: *> \endverbatim
   93: *>
   94: *> \param[in] LDB
   95: *> \verbatim
   96: *>          LDB is INTEGER
   97: *>          The leading dimension of the array B.  LDB >= max(1,N).
   98: *> \endverbatim
   99: *>
  100: *> \param[in,out] X
  101: *> \verbatim
  102: *>          X is DOUBLE PRECISION array, dimension (LDX,NRHS)
  103: *>          On entry, the solution matrix X, as computed by DPPTRS.
  104: *>          On exit, the improved solution matrix X.
  105: *> \endverbatim
  106: *>
  107: *> \param[in] LDX
  108: *> \verbatim
  109: *>          LDX is INTEGER
  110: *>          The leading dimension of the array X.  LDX >= max(1,N).
  111: *> \endverbatim
  112: *>
  113: *> \param[out] FERR
  114: *> \verbatim
  115: *>          FERR is DOUBLE PRECISION array, dimension (NRHS)
  116: *>          The estimated forward error bound for each solution vector
  117: *>          X(j) (the j-th column of the solution matrix X).
  118: *>          If XTRUE is the true solution corresponding to X(j), FERR(j)
  119: *>          is an estimated upper bound for the magnitude of the largest
  120: *>          element in (X(j) - XTRUE) divided by the magnitude of the
  121: *>          largest element in X(j).  The estimate is as reliable as
  122: *>          the estimate for RCOND, and is almost always a slight
  123: *>          overestimate of the true error.
  124: *> \endverbatim
  125: *>
  126: *> \param[out] BERR
  127: *> \verbatim
  128: *>          BERR is DOUBLE PRECISION array, dimension (NRHS)
  129: *>          The componentwise relative backward error of each solution
  130: *>          vector X(j) (i.e., the smallest relative change in
  131: *>          any element of A or B that makes X(j) an exact solution).
  132: *> \endverbatim
  133: *>
  134: *> \param[out] WORK
  135: *> \verbatim
  136: *>          WORK is DOUBLE PRECISION array, dimension (3*N)
  137: *> \endverbatim
  138: *>
  139: *> \param[out] IWORK
  140: *> \verbatim
  141: *>          IWORK is INTEGER array, dimension (N)
  142: *> \endverbatim
  143: *>
  144: *> \param[out] INFO
  145: *> \verbatim
  146: *>          INFO is INTEGER
  147: *>          = 0:  successful exit
  148: *>          < 0:  if INFO = -i, the i-th argument had an illegal value
  149: *> \endverbatim
  150: *
  151: *> \par Internal Parameters:
  152: *  =========================
  153: *>
  154: *> \verbatim
  155: *>  ITMAX is the maximum number of steps of iterative refinement.
  156: *> \endverbatim
  157: *
  158: *  Authors:
  159: *  ========
  160: *
  161: *> \author Univ. of Tennessee 
  162: *> \author Univ. of California Berkeley 
  163: *> \author Univ. of Colorado Denver 
  164: *> \author NAG Ltd. 
  165: *
  166: *> \date November 2011
  167: *
  168: *> \ingroup doubleOTHERcomputational
  169: *
  170: *  =====================================================================
  171:       SUBROUTINE DPPRFS( UPLO, N, NRHS, AP, AFP, B, LDB, X, LDX, FERR,
  172:      $                   BERR, WORK, IWORK, INFO )
  173: *
  174: *  -- LAPACK computational routine (version 3.4.0) --
  175: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  176: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  177: *     November 2011
  178: *
  179: *     .. Scalar Arguments ..
  180:       CHARACTER          UPLO
  181:       INTEGER            INFO, LDB, LDX, N, NRHS
  182: *     ..
  183: *     .. Array Arguments ..
  184:       INTEGER            IWORK( * )
  185:       DOUBLE PRECISION   AFP( * ), AP( * ), B( LDB, * ), BERR( * ),
  186:      $                   FERR( * ), WORK( * ), X( LDX, * )
  187: *     ..
  188: *
  189: *  =====================================================================
  190: *
  191: *     .. Parameters ..
  192:       INTEGER            ITMAX
  193:       PARAMETER          ( ITMAX = 5 )
  194:       DOUBLE PRECISION   ZERO
  195:       PARAMETER          ( ZERO = 0.0D+0 )
  196:       DOUBLE PRECISION   ONE
  197:       PARAMETER          ( ONE = 1.0D+0 )
  198:       DOUBLE PRECISION   TWO
  199:       PARAMETER          ( TWO = 2.0D+0 )
  200:       DOUBLE PRECISION   THREE
  201:       PARAMETER          ( THREE = 3.0D+0 )
  202: *     ..
  203: *     .. Local Scalars ..
  204:       LOGICAL            UPPER
  205:       INTEGER            COUNT, I, IK, J, K, KASE, KK, NZ
  206:       DOUBLE PRECISION   EPS, LSTRES, S, SAFE1, SAFE2, SAFMIN, XK
  207: *     ..
  208: *     .. Local Arrays ..
  209:       INTEGER            ISAVE( 3 )
  210: *     ..
  211: *     .. External Subroutines ..
  212:       EXTERNAL           DAXPY, DCOPY, DLACN2, DPPTRS, DSPMV, XERBLA
  213: *     ..
  214: *     .. Intrinsic Functions ..
  215:       INTRINSIC          ABS, MAX
  216: *     ..
  217: *     .. External Functions ..
  218:       LOGICAL            LSAME
  219:       DOUBLE PRECISION   DLAMCH
  220:       EXTERNAL           LSAME, DLAMCH
  221: *     ..
  222: *     .. Executable Statements ..
  223: *
  224: *     Test the input parameters.
  225: *
  226:       INFO = 0
  227:       UPPER = LSAME( UPLO, 'U' )
  228:       IF( .NOT.UPPER .AND. .NOT.LSAME( UPLO, 'L' ) ) THEN
  229:          INFO = -1
  230:       ELSE IF( N.LT.0 ) THEN
  231:          INFO = -2
  232:       ELSE IF( NRHS.LT.0 ) THEN
  233:          INFO = -3
  234:       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
  235:          INFO = -7
  236:       ELSE IF( LDX.LT.MAX( 1, N ) ) THEN
  237:          INFO = -9
  238:       END IF
  239:       IF( INFO.NE.0 ) THEN
  240:          CALL XERBLA( 'DPPRFS', -INFO )
  241:          RETURN
  242:       END IF
  243: *
  244: *     Quick return if possible
  245: *
  246:       IF( N.EQ.0 .OR. NRHS.EQ.0 ) THEN
  247:          DO 10 J = 1, NRHS
  248:             FERR( J ) = ZERO
  249:             BERR( J ) = ZERO
  250:    10    CONTINUE
  251:          RETURN
  252:       END IF
  253: *
  254: *     NZ = maximum number of nonzero elements in each row of A, plus 1
  255: *
  256:       NZ = N + 1
  257:       EPS = DLAMCH( 'Epsilon' )
  258:       SAFMIN = DLAMCH( 'Safe minimum' )
  259:       SAFE1 = NZ*SAFMIN
  260:       SAFE2 = SAFE1 / EPS
  261: *
  262: *     Do for each right hand side
  263: *
  264:       DO 140 J = 1, NRHS
  265: *
  266:          COUNT = 1
  267:          LSTRES = THREE
  268:    20    CONTINUE
  269: *
  270: *        Loop until stopping criterion is satisfied.
  271: *
  272: *        Compute residual R = B - A * X
  273: *
  274:          CALL DCOPY( N, B( 1, J ), 1, WORK( N+1 ), 1 )
  275:          CALL DSPMV( UPLO, N, -ONE, AP, X( 1, J ), 1, ONE, WORK( N+1 ),
  276:      $               1 )
  277: *
  278: *        Compute componentwise relative backward error from formula
  279: *
  280: *        max(i) ( abs(R(i)) / ( abs(A)*abs(X) + abs(B) )(i) )
  281: *
  282: *        where abs(Z) is the componentwise absolute value of the matrix
  283: *        or vector Z.  If the i-th component of the denominator is less
  284: *        than SAFE2, then SAFE1 is added to the i-th components of the
  285: *        numerator and denominator before dividing.
  286: *
  287:          DO 30 I = 1, N
  288:             WORK( I ) = ABS( B( I, J ) )
  289:    30    CONTINUE
  290: *
  291: *        Compute abs(A)*abs(X) + abs(B).
  292: *
  293:          KK = 1
  294:          IF( UPPER ) THEN
  295:             DO 50 K = 1, N
  296:                S = ZERO
  297:                XK = ABS( X( K, J ) )
  298:                IK = KK
  299:                DO 40 I = 1, K - 1
  300:                   WORK( I ) = WORK( I ) + ABS( AP( IK ) )*XK
  301:                   S = S + ABS( AP( IK ) )*ABS( X( I, J ) )
  302:                   IK = IK + 1
  303:    40          CONTINUE
  304:                WORK( K ) = WORK( K ) + ABS( AP( KK+K-1 ) )*XK + S
  305:                KK = KK + K
  306:    50       CONTINUE
  307:          ELSE
  308:             DO 70 K = 1, N
  309:                S = ZERO
  310:                XK = ABS( X( K, J ) )
  311:                WORK( K ) = WORK( K ) + ABS( AP( KK ) )*XK
  312:                IK = KK + 1
  313:                DO 60 I = K + 1, N
  314:                   WORK( I ) = WORK( I ) + ABS( AP( IK ) )*XK
  315:                   S = S + ABS( AP( IK ) )*ABS( X( I, J ) )
  316:                   IK = IK + 1
  317:    60          CONTINUE
  318:                WORK( K ) = WORK( K ) + S
  319:                KK = KK + ( N-K+1 )
  320:    70       CONTINUE
  321:          END IF
  322:          S = ZERO
  323:          DO 80 I = 1, N
  324:             IF( WORK( I ).GT.SAFE2 ) THEN
  325:                S = MAX( S, ABS( WORK( N+I ) ) / WORK( I ) )
  326:             ELSE
  327:                S = MAX( S, ( ABS( WORK( N+I ) )+SAFE1 ) /
  328:      $             ( WORK( I )+SAFE1 ) )
  329:             END IF
  330:    80    CONTINUE
  331:          BERR( J ) = S
  332: *
  333: *        Test stopping criterion. Continue iterating if
  334: *           1) The residual BERR(J) is larger than machine epsilon, and
  335: *           2) BERR(J) decreased by at least a factor of 2 during the
  336: *              last iteration, and
  337: *           3) At most ITMAX iterations tried.
  338: *
  339:          IF( BERR( J ).GT.EPS .AND. TWO*BERR( J ).LE.LSTRES .AND.
  340:      $       COUNT.LE.ITMAX ) THEN
  341: *
  342: *           Update solution and try again.
  343: *
  344:             CALL DPPTRS( UPLO, N, 1, AFP, WORK( N+1 ), N, INFO )
  345:             CALL DAXPY( N, ONE, WORK( N+1 ), 1, X( 1, J ), 1 )
  346:             LSTRES = BERR( J )
  347:             COUNT = COUNT + 1
  348:             GO TO 20
  349:          END IF
  350: *
  351: *        Bound error from formula
  352: *
  353: *        norm(X - XTRUE) / norm(X) .le. FERR =
  354: *        norm( abs(inv(A))*
  355: *           ( abs(R) + NZ*EPS*( abs(A)*abs(X)+abs(B) ))) / norm(X)
  356: *
  357: *        where
  358: *          norm(Z) is the magnitude of the largest component of Z
  359: *          inv(A) is the inverse of A
  360: *          abs(Z) is the componentwise absolute value of the matrix or
  361: *             vector Z
  362: *          NZ is the maximum number of nonzeros in any row of A, plus 1
  363: *          EPS is machine epsilon
  364: *
  365: *        The i-th component of abs(R)+NZ*EPS*(abs(A)*abs(X)+abs(B))
  366: *        is incremented by SAFE1 if the i-th component of
  367: *        abs(A)*abs(X) + abs(B) is less than SAFE2.
  368: *
  369: *        Use DLACN2 to estimate the infinity-norm of the matrix
  370: *           inv(A) * diag(W),
  371: *        where W = abs(R) + NZ*EPS*( abs(A)*abs(X)+abs(B) )))
  372: *
  373:          DO 90 I = 1, N
  374:             IF( WORK( I ).GT.SAFE2 ) THEN
  375:                WORK( I ) = ABS( WORK( N+I ) ) + NZ*EPS*WORK( I )
  376:             ELSE
  377:                WORK( I ) = ABS( WORK( N+I ) ) + NZ*EPS*WORK( I ) + SAFE1
  378:             END IF
  379:    90    CONTINUE
  380: *
  381:          KASE = 0
  382:   100    CONTINUE
  383:          CALL DLACN2( N, WORK( 2*N+1 ), WORK( N+1 ), IWORK, FERR( J ),
  384:      $                KASE, ISAVE )
  385:          IF( KASE.NE.0 ) THEN
  386:             IF( KASE.EQ.1 ) THEN
  387: *
  388: *              Multiply by diag(W)*inv(A**T).
  389: *
  390:                CALL DPPTRS( UPLO, N, 1, AFP, WORK( N+1 ), N, INFO )
  391:                DO 110 I = 1, N
  392:                   WORK( N+I ) = WORK( I )*WORK( N+I )
  393:   110          CONTINUE
  394:             ELSE IF( KASE.EQ.2 ) THEN
  395: *
  396: *              Multiply by inv(A)*diag(W).
  397: *
  398:                DO 120 I = 1, N
  399:                   WORK( N+I ) = WORK( I )*WORK( N+I )
  400:   120          CONTINUE
  401:                CALL DPPTRS( UPLO, N, 1, AFP, WORK( N+1 ), N, INFO )
  402:             END IF
  403:             GO TO 100
  404:          END IF
  405: *
  406: *        Normalize error.
  407: *
  408:          LSTRES = ZERO
  409:          DO 130 I = 1, N
  410:             LSTRES = MAX( LSTRES, ABS( X( I, J ) ) )
  411:   130    CONTINUE
  412:          IF( LSTRES.NE.ZERO )
  413:      $      FERR( J ) = FERR( J ) / LSTRES
  414: *
  415:   140 CONTINUE
  416: *
  417:       RETURN
  418: *
  419: *     End of DPPRFS
  420: *
  421:       END

CVSweb interface <joel.bertrand@systella.fr>