 LAPACK  3.10.1 LAPACK: Linear Algebra PACKage

## ◆ claein()

 subroutine claein ( logical RIGHTV, logical NOINIT, integer N, complex, dimension( ldh, * ) H, integer LDH, complex W, complex, dimension( * ) V, complex, dimension( ldb, * ) B, integer LDB, real, dimension( * ) RWORK, real EPS3, real SMLNUM, integer INFO )

CLAEIN computes a specified right or left eigenvector of an upper Hessenberg matrix by inverse iteration.

Purpose:
``` CLAEIN uses inverse iteration to find a right or left eigenvector
corresponding to the eigenvalue W of a complex upper Hessenberg
matrix H.```
Parameters
 [in] RIGHTV ``` RIGHTV is LOGICAL = .TRUE. : compute right eigenvector; = .FALSE.: compute left eigenvector.``` [in] NOINIT ``` NOINIT is LOGICAL = .TRUE. : no initial vector supplied in V = .FALSE.: initial vector supplied in V.``` [in] N ``` N is INTEGER The order of the matrix H. N >= 0.``` [in] H ``` H is COMPLEX array, dimension (LDH,N) The upper Hessenberg matrix H.``` [in] LDH ``` LDH is INTEGER The leading dimension of the array H. LDH >= max(1,N).``` [in] W ``` W is COMPLEX The eigenvalue of H whose corresponding right or left eigenvector is to be computed.``` [in,out] V ``` V is COMPLEX array, dimension (N) On entry, if NOINIT = .FALSE., V must contain a starting vector for inverse iteration; otherwise V need not be set. On exit, V contains the computed eigenvector, normalized so that the component of largest magnitude has magnitude 1; here the magnitude of a complex number (x,y) is taken to be |x| + |y|.``` [out] B ` B is COMPLEX array, dimension (LDB,N)` [in] LDB ``` LDB is INTEGER The leading dimension of the array B. LDB >= max(1,N).``` [out] RWORK ` RWORK is REAL array, dimension (N)` [in] EPS3 ``` EPS3 is REAL A small machine-dependent value which is used to perturb close eigenvalues, and to replace zero pivots.``` [in] SMLNUM ``` SMLNUM is REAL A machine-dependent value close to the underflow threshold.``` [out] INFO ``` INFO is INTEGER = 0: successful exit = 1: inverse iteration did not converge; V is set to the last iterate.```

Definition at line 147 of file claein.f.

149 *
150 * -- LAPACK auxiliary routine --
151 * -- LAPACK is a software package provided by Univ. of Tennessee, --
152 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
153 *
154 * .. Scalar Arguments ..
155  LOGICAL NOINIT, RIGHTV
156  INTEGER INFO, LDB, LDH, N
157  REAL EPS3, SMLNUM
158  COMPLEX W
159 * ..
160 * .. Array Arguments ..
161  REAL RWORK( * )
162  COMPLEX B( LDB, * ), H( LDH, * ), V( * )
163 * ..
164 *
165 * =====================================================================
166 *
167 * .. Parameters ..
168  REAL ONE, TENTH
169  parameter( one = 1.0e+0, tenth = 1.0e-1 )
170  COMPLEX ZERO
171  parameter( zero = ( 0.0e+0, 0.0e+0 ) )
172 * ..
173 * .. Local Scalars ..
174  CHARACTER NORMIN, TRANS
175  INTEGER I, IERR, ITS, J
176  REAL GROWTO, NRMSML, ROOTN, RTEMP, SCALE, VNORM
177  COMPLEX CDUM, EI, EJ, TEMP, X
178 * ..
179 * .. External Functions ..
180  INTEGER ICAMAX
181  REAL SCASUM, SCNRM2
183  EXTERNAL icamax, scasum, scnrm2, cladiv
184 * ..
185 * .. External Subroutines ..
186  EXTERNAL clatrs, csscal
187 * ..
188 * .. Intrinsic Functions ..
189  INTRINSIC abs, aimag, max, real, sqrt
190 * ..
191 * .. Statement Functions ..
192  REAL CABS1
193 * ..
194 * .. Statement Function definitions ..
195  cabs1( cdum ) = abs( real( cdum ) ) + abs( aimag( cdum ) )
196 * ..
197 * .. Executable Statements ..
198 *
199  info = 0
200 *
201 * GROWTO is the threshold used in the acceptance test for an
202 * eigenvector.
203 *
204  rootn = sqrt( real( n ) )
205  growto = tenth / rootn
206  nrmsml = max( one, eps3*rootn )*smlnum
207 *
208 * Form B = H - W*I (except that the subdiagonal elements are not
209 * stored).
210 *
211  DO 20 j = 1, n
212  DO 10 i = 1, j - 1
213  b( i, j ) = h( i, j )
214  10 CONTINUE
215  b( j, j ) = h( j, j ) - w
216  20 CONTINUE
217 *
218  IF( noinit ) THEN
219 *
220 * Initialize V.
221 *
222  DO 30 i = 1, n
223  v( i ) = eps3
224  30 CONTINUE
225  ELSE
226 *
227 * Scale supplied initial vector.
228 *
229  vnorm = scnrm2( n, v, 1 )
230  CALL csscal( n, ( eps3*rootn ) / max( vnorm, nrmsml ), v, 1 )
231  END IF
232 *
233  IF( rightv ) THEN
234 *
235 * LU decomposition with partial pivoting of B, replacing zero
236 * pivots by EPS3.
237 *
238  DO 60 i = 1, n - 1
239  ei = h( i+1, i )
240  IF( cabs1( b( i, i ) ).LT.cabs1( ei ) ) THEN
241 *
242 * Interchange rows and eliminate.
243 *
244  x = cladiv( b( i, i ), ei )
245  b( i, i ) = ei
246  DO 40 j = i + 1, n
247  temp = b( i+1, j )
248  b( i+1, j ) = b( i, j ) - x*temp
249  b( i, j ) = temp
250  40 CONTINUE
251  ELSE
252 *
253 * Eliminate without interchange.
254 *
255  IF( b( i, i ).EQ.zero )
256  \$ b( i, i ) = eps3
257  x = cladiv( ei, b( i, i ) )
258  IF( x.NE.zero ) THEN
259  DO 50 j = i + 1, n
260  b( i+1, j ) = b( i+1, j ) - x*b( i, j )
261  50 CONTINUE
262  END IF
263  END IF
264  60 CONTINUE
265  IF( b( n, n ).EQ.zero )
266  \$ b( n, n ) = eps3
267 *
268  trans = 'N'
269 *
270  ELSE
271 *
272 * UL decomposition with partial pivoting of B, replacing zero
273 * pivots by EPS3.
274 *
275  DO 90 j = n, 2, -1
276  ej = h( j, j-1 )
277  IF( cabs1( b( j, j ) ).LT.cabs1( ej ) ) THEN
278 *
279 * Interchange columns and eliminate.
280 *
281  x = cladiv( b( j, j ), ej )
282  b( j, j ) = ej
283  DO 70 i = 1, j - 1
284  temp = b( i, j-1 )
285  b( i, j-1 ) = b( i, j ) - x*temp
286  b( i, j ) = temp
287  70 CONTINUE
288  ELSE
289 *
290 * Eliminate without interchange.
291 *
292  IF( b( j, j ).EQ.zero )
293  \$ b( j, j ) = eps3
294  x = cladiv( ej, b( j, j ) )
295  IF( x.NE.zero ) THEN
296  DO 80 i = 1, j - 1
297  b( i, j-1 ) = b( i, j-1 ) - x*b( i, j )
298  80 CONTINUE
299  END IF
300  END IF
301  90 CONTINUE
302  IF( b( 1, 1 ).EQ.zero )
303  \$ b( 1, 1 ) = eps3
304 *
305  trans = 'C'
306 *
307  END IF
308 *
309  normin = 'N'
310  DO 110 its = 1, n
311 *
312 * Solve U*x = scale*v for a right eigenvector
313 * or U**H *x = scale*v for a left eigenvector,
314 * overwriting x on v.
315 *
316  CALL clatrs( 'Upper', trans, 'Nonunit', normin, n, b, ldb, v,
317  \$ scale, rwork, ierr )
318  normin = 'Y'
319 *
320 * Test for sufficient growth in the norm of v.
321 *
322  vnorm = scasum( n, v, 1 )
323  IF( vnorm.GE.growto*scale )
324  \$ GO TO 120
325 *
326 * Choose new orthogonal starting vector and try again.
327 *
328  rtemp = eps3 / ( rootn+one )
329  v( 1 ) = eps3
330  DO 100 i = 2, n
331  v( i ) = rtemp
332  100 CONTINUE
333  v( n-its+1 ) = v( n-its+1 ) - eps3*rootn
334  110 CONTINUE
335 *
336 * Failure to find eigenvector in N iterations.
337 *
338  info = 1
339 *
340  120 CONTINUE
341 *
342 * Normalize eigenvector.
343 *
344  i = icamax( n, v, 1 )
345  CALL csscal( n, one / cabs1( v( i ) ), v, 1 )
346 *
347  RETURN
348 *
349 * End of CLAEIN
350 *
integer function icamax(N, CX, INCX)
ICAMAX
Definition: icamax.f:71
subroutine csscal(N, SA, CX, INCX)
CSSCAL
Definition: csscal.f:78
subroutine clatrs(UPLO, TRANS, DIAG, NORMIN, N, A, LDA, X, SCALE, CNORM, INFO)
CLATRS solves a triangular system of equations with the scale factor set to prevent overflow.
Definition: clatrs.f:239