subroutine dlartgs	(	double precision	X,
		double precision	Y,
		double precision	SIGMA,
		double precision	CS,
		double precision	SN
	)

DLARTGS generates a plane rotation designed to introduce a bulge in implicit QR iteration for the bidiagonal SVD problem.

Download DLARTGS + dependencies [TGZ] [ZIP] [TXT]

Purpose:

 DLARTGS generates a plane rotation designed to introduce a bulge in
 Golub-Reinsch-style implicit QR iteration for the bidiagonal SVD
 problem. X and Y are the top-row entries, and SIGMA is the shift.
 The computed CS and SN define a plane rotation satisfying

    [  CS  SN  ]  .  [ X^2 - SIGMA ]  =  [ R ],
    [ -SN  CS  ]     [    X * Y    ]     [ 0 ]

 with R nonnegative.  If X^2 - SIGMA and X * Y are 0, then the
 rotation is by PI/2.

Parameters

[in]	X	X is DOUBLE PRECISION The (1,1) entry of an upper bidiagonal matrix.
[in]	Y	Y is DOUBLE PRECISION The (1,2) entry of an upper bidiagonal matrix.
[in]	SIGMA	SIGMA is DOUBLE PRECISION The shift.
[out]	CS	CS is DOUBLE PRECISION The cosine of the rotation.
[out]	SN	SN is DOUBLE PRECISION The sine of the rotation.

Author

Univ. of Tennessee

Univ. of California Berkeley

Univ. of Colorado Denver

NAG Ltd.

Date

September 2012

Definition at line 92 of file dlartgs.f.

*
*  -- LAPACK computational routine (version 3.4.2) --
*  -- LAPACK is a software package provided by Univ. of Tennessee,    --
*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
*     September 2012
*
*     .. Scalar Arguments ..
      DOUBLE PRECISION        cs, sigma, sn, x, y
*     ..
*
*  ===================================================================
*
*     .. Parameters ..
      DOUBLE PRECISION        negone, one, zero
      parameter                ( negone = -1.0d0, one = 1.0d0, zero = 0.0d0 )
*     ..
*     .. Local Scalars ..
      DOUBLE PRECISION        r, s, thresh, w, z
*     ..
*     .. External Functions ..
      DOUBLE PRECISION        dlamch
      EXTERNAL           dlamch
*     .. Executable Statements ..
*
      thresh = dlamch('E')
*
*     Compute the first column of B**T*B - SIGMA^2*I, up to a scale
*     factor.
*
      IF( (sigma .EQ. zero .AND. abs(x) .LT. thresh) .OR.
     $          (abs(x) .EQ. sigma .AND. y .EQ. zero) ) THEN
         z = zero
         w = zero
      ELSE IF( sigma .EQ. zero ) THEN
         IF( x .GE. zero ) THEN
            z = x
            w = y
         ELSE
            z = -x
            w = -y
         END IF
      ELSE IF( abs(x) .LT. thresh ) THEN
         z = -sigma*sigma
         w = zero
      ELSE
         IF( x .GE. zero ) THEN
            s = one
         ELSE
            s = negone
         END IF
         z = s * (abs(x)-sigma) * (s+sigma/x)
         w = s * y
      END IF
*
*     Generate the rotation.
*     CALL DLARTGP( Z, W, CS, SN, R ) might seem more natural;
*     reordering the arguments ensures that if Z = 0 then the rotation
*     is by PI/2.
*
      CALL dlartgp( w, z, sn, cs, r )
*
      RETURN
*
*     End DLARTGS
*

Here is the call graph for this function:

Here is the caller graph for this function: