Annotation of rpl/lapack/lapack/zgbequb.f, revision 1.1

1.1     ! bertrand    1:       SUBROUTINE ZGBEQUB( M, N, KL, KU, AB, LDAB, R, C, ROWCND, COLCND,
        !             2:      $                    AMAX, INFO )
        !             3: *
        !             4: *     -- LAPACK routine (version 3.2)                                 --
        !             5: *     -- Contributed by James Demmel, Deaglan Halligan, Yozo Hida and --
        !             6: *     -- Jason Riedy of Univ. of California Berkeley.                 --
        !             7: *     -- November 2008                                                --
        !             8: *
        !             9: *     -- LAPACK is a software package provided by Univ. of Tennessee, --
        !            10: *     -- Univ. of California Berkeley and NAG Ltd.                    --
        !            11: *
        !            12:       IMPLICIT NONE
        !            13: *     ..
        !            14: *     .. Scalar Arguments ..
        !            15:       INTEGER            INFO, KL, KU, LDAB, M, N
        !            16:       DOUBLE PRECISION   AMAX, COLCND, ROWCND
        !            17: *     ..
        !            18: *     .. Array Arguments ..
        !            19:       DOUBLE PRECISION   C( * ), R( * )
        !            20:       COMPLEX*16         AB( LDAB, * )
        !            21: *     ..
        !            22: *
        !            23: *  Purpose
        !            24: *  =======
        !            25: *
        !            26: *  ZGBEQUB computes row and column scalings intended to equilibrate an
        !            27: *  M-by-N matrix A and reduce its condition number.  R returns the row
        !            28: *  scale factors and C the column scale factors, chosen to try to make
        !            29: *  the largest element in each row and column of the matrix B with
        !            30: *  elements B(i,j)=R(i)*A(i,j)*C(j) have an absolute value of at most
        !            31: *  the radix.
        !            32: *
        !            33: *  R(i) and C(j) are restricted to be a power of the radix between
        !            34: *  SMLNUM = smallest safe number and BIGNUM = largest safe number.  Use
        !            35: *  of these scaling factors is not guaranteed to reduce the condition
        !            36: *  number of A but works well in practice.
        !            37: *
        !            38: *  This routine differs from ZGEEQU by restricting the scaling factors
        !            39: *  to a power of the radix.  Baring over- and underflow, scaling by
        !            40: *  these factors introduces no additional rounding errors.  However, the
        !            41: *  scaled entries' magnitured are no longer approximately 1 but lie
        !            42: *  between sqrt(radix) and 1/sqrt(radix).
        !            43: *
        !            44: *  Arguments
        !            45: *  =========
        !            46: *
        !            47: *  M       (input) INTEGER
        !            48: *          The number of rows of the matrix A.  M >= 0.
        !            49: *
        !            50: *  N       (input) INTEGER
        !            51: *          The number of columns of the matrix A.  N >= 0.
        !            52: *
        !            53: *  KL      (input) INTEGER
        !            54: *          The number of subdiagonals within the band of A.  KL >= 0.
        !            55: *
        !            56: *  KU      (input) INTEGER
        !            57: *          The number of superdiagonals within the band of A.  KU >= 0.
        !            58: *
        !            59: *  AB      (input) DOUBLE PRECISION array, dimension (LDAB,N)
        !            60: *          On entry, the matrix A in band storage, in rows 1 to KL+KU+1.
        !            61: *          The j-th column of A is stored in the j-th column of the
        !            62: *          array AB as follows:
        !            63: *          AB(KU+1+i-j,j) = A(i,j) for max(1,j-KU)<=i<=min(N,j+kl)
        !            64: *
        !            65: *  LDAB    (input) INTEGER
        !            66: *          The leading dimension of the array A.  LDAB >= max(1,M).
        !            67: *
        !            68: *  R       (output) DOUBLE PRECISION array, dimension (M)
        !            69: *          If INFO = 0 or INFO > M, R contains the row scale factors
        !            70: *          for A.
        !            71: *
        !            72: *  C       (output) DOUBLE PRECISION array, dimension (N)
        !            73: *          If INFO = 0,  C contains the column scale factors for A.
        !            74: *
        !            75: *  ROWCND  (output) DOUBLE PRECISION
        !            76: *          If INFO = 0 or INFO > M, ROWCND contains the ratio of the
        !            77: *          smallest R(i) to the largest R(i).  If ROWCND >= 0.1 and
        !            78: *          AMAX is neither too large nor too small, it is not worth
        !            79: *          scaling by R.
        !            80: *
        !            81: *  COLCND  (output) DOUBLE PRECISION
        !            82: *          If INFO = 0, COLCND contains the ratio of the smallest
        !            83: *          C(i) to the largest C(i).  If COLCND >= 0.1, it is not
        !            84: *          worth scaling by C.
        !            85: *
        !            86: *  AMAX    (output) DOUBLE PRECISION
        !            87: *          Absolute value of largest matrix element.  If AMAX is very
        !            88: *          close to overflow or very close to underflow, the matrix
        !            89: *          should be scaled.
        !            90: *
        !            91: *  INFO    (output) INTEGER
        !            92: *          = 0:  successful exit
        !            93: *          < 0:  if INFO = -i, the i-th argument had an illegal value
        !            94: *          > 0:  if INFO = i,  and i is
        !            95: *                <= M:  the i-th row of A is exactly zero
        !            96: *                >  M:  the (i-M)-th column of A is exactly zero
        !            97: *
        !            98: *  =====================================================================
        !            99: *
        !           100: *     .. Parameters ..
        !           101:       DOUBLE PRECISION   ONE, ZERO
        !           102:       PARAMETER          ( ONE = 1.0D+0, ZERO = 0.0D+0 )
        !           103: *     ..
        !           104: *     .. Local Scalars ..
        !           105:       INTEGER            I, J, KD
        !           106:       DOUBLE PRECISION   BIGNUM, RCMAX, RCMIN, SMLNUM, RADIX,
        !           107:      $                   LOGRDX
        !           108:       COMPLEX*16         ZDUM
        !           109: *     ..
        !           110: *     .. External Functions ..
        !           111:       DOUBLE PRECISION   DLAMCH
        !           112:       EXTERNAL           DLAMCH
        !           113: *     ..
        !           114: *     .. External Subroutines ..
        !           115:       EXTERNAL           XERBLA
        !           116: *     ..
        !           117: *     .. Intrinsic Functions ..
        !           118:       INTRINSIC          ABS, MAX, MIN, LOG, REAL, DIMAG
        !           119: *     ..
        !           120: *     .. Statement Functions ..
        !           121:       DOUBLE PRECISION   CABS1
        !           122: *     ..
        !           123: *     .. Statement Function definitions ..
        !           124:       CABS1( ZDUM ) = ABS( DBLE( ZDUM ) ) + ABS( DIMAG( ZDUM ) )
        !           125: *     ..
        !           126: *     .. Executable Statements ..
        !           127: *
        !           128: *     Test the input parameters.
        !           129: *
        !           130:       INFO = 0
        !           131:       IF( M.LT.0 ) THEN
        !           132:          INFO = -1
        !           133:       ELSE IF( N.LT.0 ) THEN
        !           134:          INFO = -2
        !           135:       ELSE IF( KL.LT.0 ) THEN
        !           136:          INFO = -3
        !           137:       ELSE IF( KU.LT.0 ) THEN
        !           138:          INFO = -4
        !           139:       ELSE IF( LDAB.LT.KL+KU+1 ) THEN
        !           140:          INFO = -6
        !           141:       END IF
        !           142:       IF( INFO.NE.0 ) THEN
        !           143:          CALL XERBLA( 'ZGBEQUB', -INFO )
        !           144:          RETURN
        !           145:       END IF
        !           146: *
        !           147: *     Quick return if possible.
        !           148: *
        !           149:       IF( M.EQ.0 .OR. N.EQ.0 ) THEN
        !           150:          ROWCND = ONE
        !           151:          COLCND = ONE
        !           152:          AMAX = ZERO
        !           153:          RETURN
        !           154:       END IF
        !           155: *
        !           156: *     Get machine constants.  Assume SMLNUM is a power of the radix.
        !           157: *
        !           158:       SMLNUM = DLAMCH( 'S' )
        !           159:       BIGNUM = ONE / SMLNUM
        !           160:       RADIX = DLAMCH( 'B' )
        !           161:       LOGRDX = LOG(RADIX)
        !           162: *
        !           163: *     Compute row scale factors.
        !           164: *
        !           165:       DO 10 I = 1, M
        !           166:          R( I ) = ZERO
        !           167:    10 CONTINUE
        !           168: *
        !           169: *     Find the maximum element in each row.
        !           170: *
        !           171:       KD = KU + 1
        !           172:       DO 30 J = 1, N
        !           173:          DO 20 I = MAX( J-KU, 1 ), MIN( J+KL, M )
        !           174:             R( I ) = MAX( R( I ), CABS1( AB( KD+I-J, J ) ) )
        !           175:    20    CONTINUE
        !           176:    30 CONTINUE
        !           177:       DO I = 1, M
        !           178:          IF( R( I ).GT.ZERO ) THEN
        !           179:             R( I ) = RADIX**INT( LOG( R( I ) ) / LOGRDX )
        !           180:          END IF
        !           181:       END DO
        !           182: *
        !           183: *     Find the maximum and minimum scale factors.
        !           184: *
        !           185:       RCMIN = BIGNUM
        !           186:       RCMAX = ZERO
        !           187:       DO 40 I = 1, M
        !           188:          RCMAX = MAX( RCMAX, R( I ) )
        !           189:          RCMIN = MIN( RCMIN, R( I ) )
        !           190:    40 CONTINUE
        !           191:       AMAX = RCMAX
        !           192: *
        !           193:       IF( RCMIN.EQ.ZERO ) THEN
        !           194: *
        !           195: *        Find the first zero scale factor and return an error code.
        !           196: *
        !           197:          DO 50 I = 1, M
        !           198:             IF( R( I ).EQ.ZERO ) THEN
        !           199:                INFO = I
        !           200:                RETURN
        !           201:             END IF
        !           202:    50    CONTINUE
        !           203:       ELSE
        !           204: *
        !           205: *        Invert the scale factors.
        !           206: *
        !           207:          DO 60 I = 1, M
        !           208:             R( I ) = ONE / MIN( MAX( R( I ), SMLNUM ), BIGNUM )
        !           209:    60    CONTINUE
        !           210: *
        !           211: *        Compute ROWCND = min(R(I)) / max(R(I)).
        !           212: *
        !           213:          ROWCND = MAX( RCMIN, SMLNUM ) / MIN( RCMAX, BIGNUM )
        !           214:       END IF
        !           215: *
        !           216: *     Compute column scale factors.
        !           217: *
        !           218:       DO 70 J = 1, N
        !           219:          C( J ) = ZERO
        !           220:    70 CONTINUE
        !           221: *
        !           222: *     Find the maximum element in each column,
        !           223: *     assuming the row scaling computed above.
        !           224: *
        !           225:       DO 90 J = 1, N
        !           226:          DO 80 I = MAX( J-KU, 1 ), MIN( J+KL, M )
        !           227:             C( J ) = MAX( C( J ), CABS1( AB( KD+I-J, J ) )*R( I ) )
        !           228:    80    CONTINUE
        !           229:          IF( C( J ).GT.ZERO ) THEN
        !           230:             C( J ) = RADIX**INT( LOG( C( J ) ) / LOGRDX )
        !           231:          END IF
        !           232:    90 CONTINUE
        !           233: *
        !           234: *     Find the maximum and minimum scale factors.
        !           235: *
        !           236:       RCMIN = BIGNUM
        !           237:       RCMAX = ZERO
        !           238:       DO 100 J = 1, N
        !           239:          RCMIN = MIN( RCMIN, C( J ) )
        !           240:          RCMAX = MAX( RCMAX, C( J ) )
        !           241:   100 CONTINUE
        !           242: *
        !           243:       IF( RCMIN.EQ.ZERO ) THEN
        !           244: *
        !           245: *        Find the first zero scale factor and return an error code.
        !           246: *
        !           247:          DO 110 J = 1, N
        !           248:             IF( C( J ).EQ.ZERO ) THEN
        !           249:                INFO = M + J
        !           250:                RETURN
        !           251:             END IF
        !           252:   110    CONTINUE
        !           253:       ELSE
        !           254: *
        !           255: *        Invert the scale factors.
        !           256: *
        !           257:          DO 120 J = 1, N
        !           258:             C( J ) = ONE / MIN( MAX( C( J ), SMLNUM ), BIGNUM )
        !           259:   120    CONTINUE
        !           260: *
        !           261: *        Compute COLCND = min(C(J)) / max(C(J)).
        !           262: *
        !           263:          COLCND = MAX( RCMIN, SMLNUM ) / MIN( RCMAX, BIGNUM )
        !           264:       END IF
        !           265: *
        !           266:       RETURN
        !           267: *
        !           268: *     End of ZGBEQUB
        !           269: *
        !           270:       END

CVSweb interface <joel.bertrand@systella.fr>