Annotation of rpl/lapack/lapack/dggrqf.f, revision 1.18
1.9 bertrand 1: *> \brief \b DGGRQF
2: *
3: * =========== DOCUMENTATION ===========
4: *
1.15 bertrand 5: * Online html documentation available at
6: * http://www.netlib.org/lapack/explore-html/
1.9 bertrand 7: *
8: *> \htmlonly
1.15 bertrand 9: *> Download DGGRQF + dependencies
10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dggrqf.f">
11: *> [TGZ]</a>
12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dggrqf.f">
13: *> [ZIP]</a>
14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dggrqf.f">
1.9 bertrand 15: *> [TXT]</a>
1.15 bertrand 16: *> \endhtmlonly
1.9 bertrand 17: *
18: * Definition:
19: * ===========
20: *
21: * SUBROUTINE DGGRQF( M, P, N, A, LDA, TAUA, B, LDB, TAUB, WORK,
22: * LWORK, INFO )
1.15 bertrand 23: *
1.9 bertrand 24: * .. Scalar Arguments ..
25: * INTEGER INFO, LDA, LDB, LWORK, M, N, P
26: * ..
27: * .. Array Arguments ..
28: * DOUBLE PRECISION A( LDA, * ), B( LDB, * ), TAUA( * ), TAUB( * ),
29: * $ WORK( * )
30: * ..
1.15 bertrand 31: *
1.9 bertrand 32: *
33: *> \par Purpose:
34: * =============
35: *>
36: *> \verbatim
37: *>
38: *> DGGRQF computes a generalized RQ factorization of an M-by-N matrix A
39: *> and a P-by-N matrix B:
40: *>
41: *> A = R*Q, B = Z*T*Q,
42: *>
43: *> where Q is an N-by-N orthogonal matrix, Z is a P-by-P orthogonal
44: *> matrix, and R and T assume one of the forms:
45: *>
46: *> if M <= N, R = ( 0 R12 ) M, or if M > N, R = ( R11 ) M-N,
47: *> N-M M ( R21 ) N
48: *> N
49: *>
50: *> where R12 or R21 is upper triangular, and
51: *>
52: *> if P >= N, T = ( T11 ) N , or if P < N, T = ( T11 T12 ) P,
53: *> ( 0 ) P-N P N-P
54: *> N
55: *>
56: *> where T11 is upper triangular.
57: *>
58: *> In particular, if B is square and nonsingular, the GRQ factorization
59: *> of A and B implicitly gives the RQ factorization of A*inv(B):
60: *>
61: *> A*inv(B) = (R*inv(T))*Z**T
62: *>
63: *> where inv(B) denotes the inverse of the matrix B, and Z**T denotes the
64: *> transpose of the matrix Z.
65: *> \endverbatim
66: *
67: * Arguments:
68: * ==========
69: *
70: *> \param[in] M
71: *> \verbatim
72: *> M is INTEGER
73: *> The number of rows of the matrix A. M >= 0.
74: *> \endverbatim
75: *>
76: *> \param[in] P
77: *> \verbatim
78: *> P is INTEGER
79: *> The number of rows of the matrix B. P >= 0.
80: *> \endverbatim
81: *>
82: *> \param[in] N
83: *> \verbatim
84: *> N is INTEGER
85: *> The number of columns of the matrices A and B. N >= 0.
86: *> \endverbatim
87: *>
88: *> \param[in,out] A
89: *> \verbatim
90: *> A is DOUBLE PRECISION array, dimension (LDA,N)
91: *> On entry, the M-by-N matrix A.
92: *> On exit, if M <= N, the upper triangle of the subarray
93: *> A(1:M,N-M+1:N) contains the M-by-M upper triangular matrix R;
94: *> if M > N, the elements on and above the (M-N)-th subdiagonal
95: *> contain the M-by-N upper trapezoidal matrix R; the remaining
96: *> elements, with the array TAUA, represent the orthogonal
97: *> matrix Q as a product of elementary reflectors (see Further
98: *> Details).
99: *> \endverbatim
100: *>
101: *> \param[in] LDA
102: *> \verbatim
103: *> LDA is INTEGER
104: *> The leading dimension of the array A. LDA >= max(1,M).
105: *> \endverbatim
106: *>
107: *> \param[out] TAUA
108: *> \verbatim
109: *> TAUA is DOUBLE PRECISION array, dimension (min(M,N))
110: *> The scalar factors of the elementary reflectors which
111: *> represent the orthogonal matrix Q (see Further Details).
112: *> \endverbatim
113: *>
114: *> \param[in,out] B
115: *> \verbatim
116: *> B is DOUBLE PRECISION array, dimension (LDB,N)
117: *> On entry, the P-by-N matrix B.
118: *> On exit, the elements on and above the diagonal of the array
119: *> contain the min(P,N)-by-N upper trapezoidal matrix T (T is
120: *> upper triangular if P >= N); the elements below the diagonal,
121: *> with the array TAUB, represent the orthogonal matrix Z as a
122: *> product of elementary reflectors (see Further Details).
123: *> \endverbatim
124: *>
125: *> \param[in] LDB
126: *> \verbatim
127: *> LDB is INTEGER
128: *> The leading dimension of the array B. LDB >= max(1,P).
129: *> \endverbatim
130: *>
131: *> \param[out] TAUB
132: *> \verbatim
133: *> TAUB is DOUBLE PRECISION array, dimension (min(P,N))
134: *> The scalar factors of the elementary reflectors which
135: *> represent the orthogonal matrix Z (see Further Details).
136: *> \endverbatim
137: *>
138: *> \param[out] WORK
139: *> \verbatim
140: *> WORK is DOUBLE PRECISION array, dimension (MAX(1,LWORK))
141: *> On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
142: *> \endverbatim
143: *>
144: *> \param[in] LWORK
145: *> \verbatim
146: *> LWORK is INTEGER
147: *> The dimension of the array WORK. LWORK >= max(1,N,M,P).
148: *> For optimum performance LWORK >= max(N,M,P)*max(NB1,NB2,NB3),
149: *> where NB1 is the optimal blocksize for the RQ factorization
150: *> of an M-by-N matrix, NB2 is the optimal blocksize for the
151: *> QR factorization of a P-by-N matrix, and NB3 is the optimal
152: *> blocksize for a call of DORMRQ.
153: *>
154: *> If LWORK = -1, then a workspace query is assumed; the routine
155: *> only calculates the optimal size of the WORK array, returns
156: *> this value as the first entry of the WORK array, and no error
157: *> message related to LWORK is issued by XERBLA.
158: *> \endverbatim
159: *>
160: *> \param[out] INFO
161: *> \verbatim
162: *> INFO is INTEGER
163: *> = 0: successful exit
164: *> < 0: if INF0= -i, the i-th argument had an illegal value.
165: *> \endverbatim
166: *
167: * Authors:
168: * ========
169: *
1.15 bertrand 170: *> \author Univ. of Tennessee
171: *> \author Univ. of California Berkeley
172: *> \author Univ. of Colorado Denver
173: *> \author NAG Ltd.
1.9 bertrand 174: *
175: *> \ingroup doubleOTHERcomputational
176: *
177: *> \par Further Details:
178: * =====================
179: *>
180: *> \verbatim
181: *>
182: *> The matrix Q is represented as a product of elementary reflectors
183: *>
184: *> Q = H(1) H(2) . . . H(k), where k = min(m,n).
185: *>
186: *> Each H(i) has the form
187: *>
188: *> H(i) = I - taua * v * v**T
189: *>
190: *> where taua is a real scalar, and v is a real vector with
191: *> v(n-k+i+1:n) = 0 and v(n-k+i) = 1; v(1:n-k+i-1) is stored on exit in
192: *> A(m-k+i,1:n-k+i-1), and taua in TAUA(i).
193: *> To form Q explicitly, use LAPACK subroutine DORGRQ.
194: *> To use Q to update another matrix, use LAPACK subroutine DORMRQ.
195: *>
196: *> The matrix Z is represented as a product of elementary reflectors
197: *>
198: *> Z = H(1) H(2) . . . H(k), where k = min(p,n).
199: *>
200: *> Each H(i) has the form
201: *>
202: *> H(i) = I - taub * v * v**T
203: *>
204: *> where taub is a real scalar, and v is a real vector with
205: *> v(1:i-1) = 0 and v(i) = 1; v(i+1:p) is stored on exit in B(i+1:p,i),
206: *> and taub in TAUB(i).
207: *> To form Z explicitly, use LAPACK subroutine DORGQR.
208: *> To use Z to update another matrix, use LAPACK subroutine DORMQR.
209: *> \endverbatim
210: *>
211: * =====================================================================
1.1 bertrand 212: SUBROUTINE DGGRQF( M, P, N, A, LDA, TAUA, B, LDB, TAUB, WORK,
213: $ LWORK, INFO )
214: *
1.18 ! bertrand 215: * -- LAPACK computational routine --
1.1 bertrand 216: * -- LAPACK is a software package provided by Univ. of Tennessee, --
217: * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
218: *
219: * .. Scalar Arguments ..
220: INTEGER INFO, LDA, LDB, LWORK, M, N, P
221: * ..
222: * .. Array Arguments ..
223: DOUBLE PRECISION A( LDA, * ), B( LDB, * ), TAUA( * ), TAUB( * ),
224: $ WORK( * )
225: * ..
226: *
227: * =====================================================================
228: *
229: * .. Local Scalars ..
230: LOGICAL LQUERY
231: INTEGER LOPT, LWKOPT, NB, NB1, NB2, NB3
232: * ..
233: * .. External Subroutines ..
234: EXTERNAL DGEQRF, DGERQF, DORMRQ, XERBLA
235: * ..
236: * .. External Functions ..
237: INTEGER ILAENV
238: EXTERNAL ILAENV
239: * ..
240: * .. Intrinsic Functions ..
241: INTRINSIC INT, MAX, MIN
242: * ..
243: * .. Executable Statements ..
244: *
245: * Test the input parameters
246: *
247: INFO = 0
248: NB1 = ILAENV( 1, 'DGERQF', ' ', M, N, -1, -1 )
249: NB2 = ILAENV( 1, 'DGEQRF', ' ', P, N, -1, -1 )
250: NB3 = ILAENV( 1, 'DORMRQ', ' ', M, N, P, -1 )
251: NB = MAX( NB1, NB2, NB3 )
252: LWKOPT = MAX( N, M, P )*NB
253: WORK( 1 ) = LWKOPT
254: LQUERY = ( LWORK.EQ.-1 )
255: IF( M.LT.0 ) THEN
256: INFO = -1
257: ELSE IF( P.LT.0 ) THEN
258: INFO = -2
259: ELSE IF( N.LT.0 ) THEN
260: INFO = -3
261: ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
262: INFO = -5
263: ELSE IF( LDB.LT.MAX( 1, P ) ) THEN
264: INFO = -8
265: ELSE IF( LWORK.LT.MAX( 1, M, P, N ) .AND. .NOT.LQUERY ) THEN
266: INFO = -11
267: END IF
268: IF( INFO.NE.0 ) THEN
269: CALL XERBLA( 'DGGRQF', -INFO )
270: RETURN
271: ELSE IF( LQUERY ) THEN
272: RETURN
273: END IF
274: *
275: * RQ factorization of M-by-N matrix A: A = R*Q
276: *
277: CALL DGERQF( M, N, A, LDA, TAUA, WORK, LWORK, INFO )
1.18 ! bertrand 278: LOPT = INT( WORK( 1 ) )
1.1 bertrand 279: *
1.8 bertrand 280: * Update B := B*Q**T
1.1 bertrand 281: *
282: CALL DORMRQ( 'Right', 'Transpose', P, N, MIN( M, N ),
283: $ A( MAX( 1, M-N+1 ), 1 ), LDA, TAUA, B, LDB, WORK,
284: $ LWORK, INFO )
285: LOPT = MAX( LOPT, INT( WORK( 1 ) ) )
286: *
287: * QR factorization of P-by-N matrix B: B = Z*T
288: *
289: CALL DGEQRF( P, N, B, LDB, TAUB, WORK, LWORK, INFO )
290: WORK( 1 ) = MAX( LOPT, INT( WORK( 1 ) ) )
291: *
292: RETURN
293: *
294: * End of DGGRQF
295: *
296: END
CVSweb interface <joel.bertrand@systella.fr>