LAPACK  3.10.0
LAPACK: Linear Algebra PACKage

◆ slaev2()

subroutine slaev2 ( real  A,
real  B,
real  C,
real  RT1,
real  RT2,
real  CS1,
real  SN1 
)

SLAEV2 computes the eigenvalues and eigenvectors of a 2-by-2 symmetric/Hermitian matrix.

Download SLAEV2 + dependencies [TGZ] [ZIP] [TXT]

Purpose:
 SLAEV2 computes the eigendecomposition of a 2-by-2 symmetric matrix
    [  A   B  ]
    [  B   C  ].
 On return, RT1 is the eigenvalue of larger absolute value, RT2 is the
 eigenvalue of smaller absolute value, and (CS1,SN1) is the unit right
 eigenvector for RT1, giving the decomposition

    [ CS1  SN1 ] [  A   B  ] [ CS1 -SN1 ]  =  [ RT1  0  ]
    [-SN1  CS1 ] [  B   C  ] [ SN1  CS1 ]     [  0  RT2 ].
Parameters
[in]A
          A is REAL
          The (1,1) element of the 2-by-2 matrix.
[in]B
          B is REAL
          The (1,2) element and the conjugate of the (2,1) element of
          the 2-by-2 matrix.
[in]C
          C is REAL
          The (2,2) element of the 2-by-2 matrix.
[out]RT1
          RT1 is REAL
          The eigenvalue of larger absolute value.
[out]RT2
          RT2 is REAL
          The eigenvalue of smaller absolute value.
[out]CS1
          CS1 is REAL
[out]SN1
          SN1 is REAL
          The vector (CS1, SN1) is a unit right eigenvector for RT1.
Author
Univ. of Tennessee
Univ. of California Berkeley
Univ. of Colorado Denver
NAG Ltd.
Further Details:
  RT1 is accurate to a few ulps barring over/underflow.

  RT2 may be inaccurate if there is massive cancellation in the
  determinant A*C-B*B; higher precision or correctly rounded or
  correctly truncated arithmetic would be needed to compute RT2
  accurately in all cases.

  CS1 and SN1 are accurate to a few ulps barring over/underflow.

  Overflow is possible only if RT1 is within a factor of 5 of overflow.
  Underflow is harmless if the input data is 0 or exceeds
     underflow_threshold / macheps.

Definition at line 119 of file slaev2.f.

120 *
121 * -- LAPACK auxiliary routine --
122 * -- LAPACK is a software package provided by Univ. of Tennessee, --
123 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
124 *
125 * .. Scalar Arguments ..
126  REAL A, B, C, CS1, RT1, RT2, SN1
127 * ..
128 *
129 * =====================================================================
130 *
131 * .. Parameters ..
132  REAL ONE
133  parameter( one = 1.0e0 )
134  REAL TWO
135  parameter( two = 2.0e0 )
136  REAL ZERO
137  parameter( zero = 0.0e0 )
138  REAL HALF
139  parameter( half = 0.5e0 )
140 * ..
141 * .. Local Scalars ..
142  INTEGER SGN1, SGN2
143  REAL AB, ACMN, ACMX, ACS, ADF, CS, CT, DF, RT, SM,
144  $ TB, TN
145 * ..
146 * .. Intrinsic Functions ..
147  INTRINSIC abs, sqrt
148 * ..
149 * .. Executable Statements ..
150 *
151 * Compute the eigenvalues
152 *
153  sm = a + c
154  df = a - c
155  adf = abs( df )
156  tb = b + b
157  ab = abs( tb )
158  IF( abs( a ).GT.abs( c ) ) THEN
159  acmx = a
160  acmn = c
161  ELSE
162  acmx = c
163  acmn = a
164  END IF
165  IF( adf.GT.ab ) THEN
166  rt = adf*sqrt( one+( ab / adf )**2 )
167  ELSE IF( adf.LT.ab ) THEN
168  rt = ab*sqrt( one+( adf / ab )**2 )
169  ELSE
170 *
171 * Includes case AB=ADF=0
172 *
173  rt = ab*sqrt( two )
174  END IF
175  IF( sm.LT.zero ) THEN
176  rt1 = half*( sm-rt )
177  sgn1 = -1
178 *
179 * Order of execution important.
180 * To get fully accurate smaller eigenvalue,
181 * next line needs to be executed in higher precision.
182 *
183  rt2 = ( acmx / rt1 )*acmn - ( b / rt1 )*b
184  ELSE IF( sm.GT.zero ) THEN
185  rt1 = half*( sm+rt )
186  sgn1 = 1
187 *
188 * Order of execution important.
189 * To get fully accurate smaller eigenvalue,
190 * next line needs to be executed in higher precision.
191 *
192  rt2 = ( acmx / rt1 )*acmn - ( b / rt1 )*b
193  ELSE
194 *
195 * Includes case RT1 = RT2 = 0
196 *
197  rt1 = half*rt
198  rt2 = -half*rt
199  sgn1 = 1
200  END IF
201 *
202 * Compute the eigenvector
203 *
204  IF( df.GE.zero ) THEN
205  cs = df + rt
206  sgn2 = 1
207  ELSE
208  cs = df - rt
209  sgn2 = -1
210  END IF
211  acs = abs( cs )
212  IF( acs.GT.ab ) THEN
213  ct = -tb / cs
214  sn1 = one / sqrt( one+ct*ct )
215  cs1 = ct*sn1
216  ELSE
217  IF( ab.EQ.zero ) THEN
218  cs1 = one
219  sn1 = zero
220  ELSE
221  tn = -cs / tb
222  cs1 = one / sqrt( one+tn*tn )
223  sn1 = tn*cs1
224  END IF
225  END IF
226  IF( sgn1.EQ.sgn2 ) THEN
227  tn = cs1
228  cs1 = -sn1
229  sn1 = tn
230  END IF
231  RETURN
232 *
233 * End of SLAEV2
234 *
Here is the caller graph for this function: