LAPACK 3.12.1
LAPACK: Linear Algebra PACKage
All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Modules Pages
zhetrf.f
Go to the documentation of this file.
1*> \brief \b ZHETRF
2*
3* =========== DOCUMENTATION ===========
4*
5* Online html documentation available at
6* http://www.netlib.org/lapack/explore-html/
7*
8*> Download ZHETRF + dependencies
9*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/zhetrf.f">
10*> [TGZ]</a>
11*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/zhetrf.f">
12*> [ZIP]</a>
13*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/zhetrf.f">
14*> [TXT]</a>
15*
16* Definition:
17* ===========
18*
19* SUBROUTINE ZHETRF( UPLO, N, A, LDA, IPIV, WORK, LWORK, INFO )
20*
21* .. Scalar Arguments ..
22* CHARACTER UPLO
23* INTEGER INFO, LDA, LWORK, N
24* ..
25* .. Array Arguments ..
26* INTEGER IPIV( * )
27* COMPLEX*16 A( LDA, * ), WORK( * )
28* ..
29*
30*
31*> \par Purpose:
32* =============
33*>
34*> \verbatim
35*>
36*> ZHETRF computes the factorization of a complex Hermitian matrix A
37*> using the Bunch-Kaufman diagonal pivoting method. The form of the
38*> factorization is
39*>
40*> A = U*D*U**H or A = L*D*L**H
41*>
42*> where U (or L) is a product of permutation and unit upper (lower)
43*> triangular matrices, and D is Hermitian and block diagonal with
44*> 1-by-1 and 2-by-2 diagonal blocks.
45*>
46*> This is the blocked version of the algorithm, calling Level 3 BLAS.
47*> \endverbatim
48*
49* Arguments:
50* ==========
51*
52*> \param[in] UPLO
53*> \verbatim
54*> UPLO is CHARACTER*1
55*> = 'U': Upper triangle of A is stored;
56*> = 'L': Lower triangle of A is stored.
57*> \endverbatim
58*>
59*> \param[in] N
60*> \verbatim
61*> N is INTEGER
62*> The order of the matrix A. N >= 0.
63*> \endverbatim
64*>
65*> \param[in,out] A
66*> \verbatim
67*> A is COMPLEX*16 array, dimension (LDA,N)
68*> On entry, the Hermitian matrix A. If UPLO = 'U', the leading
69*> N-by-N upper triangular part of A contains the upper
70*> triangular part of the matrix A, and the strictly lower
71*> triangular part of A is not referenced. If UPLO = 'L', the
72*> leading N-by-N lower triangular part of A contains the lower
73*> triangular part of the matrix A, and the strictly upper
74*> triangular part of A is not referenced.
75*>
76*> On exit, the block diagonal matrix D and the multipliers used
77*> to obtain the factor U or L (see below for further details).
78*> \endverbatim
79*>
80*> \param[in] LDA
81*> \verbatim
82*> LDA is INTEGER
83*> The leading dimension of the array A. LDA >= max(1,N).
84*> \endverbatim
85*>
86*> \param[out] IPIV
87*> \verbatim
88*> IPIV is INTEGER array, dimension (N)
89*> Details of the interchanges and the block structure of D.
90*> If IPIV(k) > 0, then rows and columns k and IPIV(k) were
91*> interchanged and D(k,k) is a 1-by-1 diagonal block.
92*> If UPLO = 'U' and IPIV(k) = IPIV(k-1) < 0, then rows and
93*> columns k-1 and -IPIV(k) were interchanged and D(k-1:k,k-1:k)
94*> is a 2-by-2 diagonal block. If UPLO = 'L' and IPIV(k) =
95*> IPIV(k+1) < 0, then rows and columns k+1 and -IPIV(k) were
96*> interchanged and D(k:k+1,k:k+1) is a 2-by-2 diagonal block.
97*> \endverbatim
98*>
99*> \param[out] WORK
100*> \verbatim
101*> WORK is COMPLEX*16 array, dimension (MAX(1,LWORK))
102*> On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
103*> \endverbatim
104*>
105*> \param[in] LWORK
106*> \verbatim
107*> LWORK is INTEGER
108*> The length of WORK. LWORK >= 1. For best performance
109*> LWORK >= N*NB, where NB is the block size returned by ILAENV.
110*> \endverbatim
111*>
112*> \param[out] INFO
113*> \verbatim
114*> INFO is INTEGER
115*> = 0: successful exit
116*> < 0: if INFO = -i, the i-th argument had an illegal value
117*> > 0: if INFO = i, D(i,i) is exactly zero. The factorization
118*> has been completed, but the block diagonal matrix D is
119*> exactly singular, and division by zero will occur if it
120*> is used to solve a system of equations.
121*> \endverbatim
122*
123* Authors:
124* ========
125*
126*> \author Univ. of Tennessee
127*> \author Univ. of California Berkeley
128*> \author Univ. of Colorado Denver
129*> \author NAG Ltd.
130*
131*> \ingroup hetrf
132*
133*> \par Further Details:
134* =====================
135*>
136*> \verbatim
137*>
138*> If UPLO = 'U', then A = U*D*U**H, where
139*> U = P(n)*U(n)* ... *P(k)U(k)* ...,
140*> i.e., U is a product of terms P(k)*U(k), where k decreases from n to
141*> 1 in steps of 1 or 2, and D is a block diagonal matrix with 1-by-1
142*> and 2-by-2 diagonal blocks D(k). P(k) is a permutation matrix as
143*> defined by IPIV(k), and U(k) is a unit upper triangular matrix, such
144*> that if the diagonal block D(k) is of order s (s = 1 or 2), then
145*>
146*> ( I v 0 ) k-s
147*> U(k) = ( 0 I 0 ) s
148*> ( 0 0 I ) n-k
149*> k-s s n-k
150*>
151*> If s = 1, D(k) overwrites A(k,k), and v overwrites A(1:k-1,k).
152*> If s = 2, the upper triangle of D(k) overwrites A(k-1,k-1), A(k-1,k),
153*> and A(k,k), and v overwrites A(1:k-2,k-1:k).
154*>
155*> If UPLO = 'L', then A = L*D*L**H, where
156*> L = P(1)*L(1)* ... *P(k)*L(k)* ...,
157*> i.e., L is a product of terms P(k)*L(k), where k increases from 1 to
158*> n in steps of 1 or 2, and D is a block diagonal matrix with 1-by-1
159*> and 2-by-2 diagonal blocks D(k). P(k) is a permutation matrix as
160*> defined by IPIV(k), and L(k) is a unit lower triangular matrix, such
161*> that if the diagonal block D(k) is of order s (s = 1 or 2), then
162*>
163*> ( I 0 0 ) k-1
164*> L(k) = ( 0 I 0 ) s
165*> ( 0 v I ) n-k-s+1
166*> k-1 s n-k-s+1
167*>
168*> If s = 1, D(k) overwrites A(k,k), and v overwrites A(k+1:n,k).
169*> If s = 2, the lower triangle of D(k) overwrites A(k,k), A(k+1,k),
170*> and A(k+1,k+1), and v overwrites A(k+2:n,k:k+1).
171*> \endverbatim
172*>
173* =====================================================================
174 SUBROUTINE zhetrf( UPLO, N, A, LDA, IPIV, WORK, LWORK, INFO )
175*
176* -- LAPACK computational routine --
177* -- LAPACK is a software package provided by Univ. of Tennessee, --
178* -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
179*
180* .. Scalar Arguments ..
181 CHARACTER UPLO
182 INTEGER INFO, LDA, LWORK, N
183* ..
184* .. Array Arguments ..
185 INTEGER IPIV( * )
186 COMPLEX*16 A( LDA, * ), WORK( * )
187* ..
188*
189* =====================================================================
190*
191* .. Local Scalars ..
192 LOGICAL LQUERY, UPPER
193 INTEGER IINFO, IWS, J, K, KB, LDWORK, LWKOPT, NB, NBMIN
194* ..
195* .. External Functions ..
196 LOGICAL LSAME
197 INTEGER ILAENV
198 EXTERNAL lsame, ilaenv
199* ..
200* .. External Subroutines ..
201 EXTERNAL xerbla, zhetf2, zlahef
202* ..
203* .. Intrinsic Functions ..
204 INTRINSIC max
205* ..
206* .. Executable Statements ..
207*
208* Test the input parameters.
209*
210 info = 0
211 upper = lsame( uplo, 'U' )
212 lquery = ( lwork.EQ.-1 )
213 IF( .NOT.upper .AND. .NOT.lsame( uplo, 'L' ) ) THEN
214 info = -1
215 ELSE IF( n.LT.0 ) THEN
216 info = -2
217 ELSE IF( lda.LT.max( 1, n ) ) THEN
218 info = -4
219 ELSE IF( lwork.LT.1 .AND. .NOT.lquery ) THEN
220 info = -7
221 END IF
222*
223 IF( info.EQ.0 ) THEN
224*
225* Determine the block size
226*
227 nb = ilaenv( 1, 'ZHETRF', uplo, n, -1, -1, -1 )
228 lwkopt = max( 1, n*nb )
229 work( 1 ) = lwkopt
230 END IF
231*
232 IF( info.NE.0 ) THEN
233 CALL xerbla( 'ZHETRF', -info )
234 RETURN
235 ELSE IF( lquery ) THEN
236 RETURN
237 END IF
238*
239 nbmin = 2
240 ldwork = n
241 IF( nb.GT.1 .AND. nb.LT.n ) THEN
242 iws = ldwork*nb
243 IF( lwork.LT.iws ) THEN
244 nb = max( lwork / ldwork, 1 )
245 nbmin = max( 2, ilaenv( 2, 'ZHETRF', uplo, n, -1, -1,
246 $ -1 ) )
247 END IF
248 ELSE
249 iws = 1
250 END IF
251 IF( nb.LT.nbmin )
252 $ nb = n
253*
254 IF( upper ) THEN
255*
256* Factorize A as U*D*U**H using the upper triangle of A
257*
258* K is the main loop index, decreasing from N to 1 in steps of
259* KB, where KB is the number of columns factorized by ZLAHEF;
260* KB is either NB or NB-1, or K for the last block
261*
262 k = n
263 10 CONTINUE
264*
265* If K < 1, exit from loop
266*
267 IF( k.LT.1 )
268 $ GO TO 40
269*
270 IF( k.GT.nb ) THEN
271*
272* Factorize columns k-kb+1:k of A and use blocked code to
273* update columns 1:k-kb
274*
275 CALL zlahef( uplo, k, nb, kb, a, lda, ipiv, work, n,
276 $ iinfo )
277 ELSE
278*
279* Use unblocked code to factorize columns 1:k of A
280*
281 CALL zhetf2( uplo, k, a, lda, ipiv, iinfo )
282 kb = k
283 END IF
284*
285* Set INFO on the first occurrence of a zero pivot
286*
287 IF( info.EQ.0 .AND. iinfo.GT.0 )
288 $ info = iinfo
289*
290* Decrease K and return to the start of the main loop
291*
292 k = k - kb
293 GO TO 10
294*
295 ELSE
296*
297* Factorize A as L*D*L**H using the lower triangle of A
298*
299* K is the main loop index, increasing from 1 to N in steps of
300* KB, where KB is the number of columns factorized by ZLAHEF;
301* KB is either NB or NB-1, or N-K+1 for the last block
302*
303 k = 1
304 20 CONTINUE
305*
306* If K > N, exit from loop
307*
308 IF( k.GT.n )
309 $ GO TO 40
310*
311 IF( k.LE.n-nb ) THEN
312*
313* Factorize columns k:k+kb-1 of A and use blocked code to
314* update columns k+kb:n
315*
316 CALL zlahef( uplo, n-k+1, nb, kb, a( k, k ), lda,
317 $ ipiv( k ),
318 $ work, n, iinfo )
319 ELSE
320*
321* Use unblocked code to factorize columns k:n of A
322*
323 CALL zhetf2( uplo, n-k+1, a( k, k ), lda, ipiv( k ),
324 $ iinfo )
325 kb = n - k + 1
326 END IF
327*
328* Set INFO on the first occurrence of a zero pivot
329*
330 IF( info.EQ.0 .AND. iinfo.GT.0 )
331 $ info = iinfo + k - 1
332*
333* Adjust IPIV
334*
335 DO 30 j = k, k + kb - 1
336 IF( ipiv( j ).GT.0 ) THEN
337 ipiv( j ) = ipiv( j ) + k - 1
338 ELSE
339 ipiv( j ) = ipiv( j ) - k + 1
340 END IF
341 30 CONTINUE
342*
343* Increase K and return to the start of the main loop
344*
345 k = k + kb
346 GO TO 20
347*
348 END IF
349*
350 40 CONTINUE
351*
352 work( 1 ) = lwkopt
353 RETURN
354*
355* End of ZHETRF
356*
357 END
subroutine xerbla(srname, info)
Definition cblat2.f:3285
subroutine zhetf2(uplo, n, a, lda, ipiv, info)
ZHETF2 computes the factorization of a complex Hermitian matrix, using the diagonal pivoting method (...
Definition zhetf2.f:189
subroutine zhetrf(uplo, n, a, lda, ipiv, work, lwork, info)
ZHETRF
Definition zhetrf.f:175
subroutine zlahef(uplo, n, nb, kb, a, lda, ipiv, w, ldw, info)
ZLAHEF computes a partial factorization of a complex Hermitian indefinite matrix using the Bunch-Kauf...
Definition zlahef.f:176