Annotation of rpl/lapack/lapack/dgetrf2.f, revision 1.5
1.1 bertrand 1: *> \brief \b DGETRF2
2: *
3: * =========== DOCUMENTATION ===========
4: *
1.4 bertrand 5: * Online html documentation available at
6: * http://www.netlib.org/lapack/explore-html/
1.1 bertrand 7: *
8: * Definition:
9: * ===========
10: *
11: * RECURSIVE SUBROUTINE DGETRF2( M, N, A, LDA, IPIV, INFO )
1.4 bertrand 12: *
1.1 bertrand 13: * .. Scalar Arguments ..
14: * INTEGER INFO, LDA, M, N
15: * ..
16: * .. Array Arguments ..
17: * INTEGER IPIV( * )
18: * DOUBLE PRECISION A( LDA, * )
19: * ..
1.4 bertrand 20: *
1.1 bertrand 21: *
22: *> \par Purpose:
23: * =============
24: *>
25: *> \verbatim
26: *>
27: *> DGETRF2 computes an LU factorization of a general M-by-N matrix A
28: *> using partial pivoting with row interchanges.
29: *>
30: *> The factorization has the form
31: *> A = P * L * U
32: *> where P is a permutation matrix, L is lower triangular with unit
33: *> diagonal elements (lower trapezoidal if m > n), and U is upper
34: *> triangular (upper trapezoidal if m < n).
35: *>
36: *> This is the recursive version of the algorithm. It divides
37: *> the matrix into four submatrices:
1.4 bertrand 38: *>
1.1 bertrand 39: *> [ A11 | A12 ] where A11 is n1 by n1 and A22 is n2 by n2
1.2 bertrand 40: *> A = [ -----|----- ] with n1 = min(m,n)/2
1.1 bertrand 41: *> [ A21 | A22 ] n2 = n-n1
1.4 bertrand 42: *>
1.1 bertrand 43: *> [ A11 ]
44: *> The subroutine calls itself to factor [ --- ],
45: *> [ A12 ]
46: *> [ A12 ]
47: *> do the swaps on [ --- ], solve A12, update A22,
48: *> [ A22 ]
49: *>
50: *> then calls itself to factor A22 and do the swaps on A21.
51: *>
52: *> \endverbatim
53: *
54: * Arguments:
55: * ==========
56: *
57: *> \param[in] M
58: *> \verbatim
59: *> M is INTEGER
60: *> The number of rows of the matrix A. M >= 0.
61: *> \endverbatim
62: *>
63: *> \param[in] N
64: *> \verbatim
65: *> N is INTEGER
66: *> The number of columns of the matrix A. N >= 0.
67: *> \endverbatim
68: *>
69: *> \param[in,out] A
70: *> \verbatim
71: *> A is DOUBLE PRECISION array, dimension (LDA,N)
72: *> On entry, the M-by-N matrix to be factored.
73: *> On exit, the factors L and U from the factorization
74: *> A = P*L*U; the unit diagonal elements of L are not stored.
75: *> \endverbatim
76: *>
77: *> \param[in] LDA
78: *> \verbatim
79: *> LDA is INTEGER
80: *> The leading dimension of the array A. LDA >= max(1,M).
81: *> \endverbatim
82: *>
83: *> \param[out] IPIV
84: *> \verbatim
85: *> IPIV is INTEGER array, dimension (min(M,N))
86: *> The pivot indices; for 1 <= i <= min(M,N), row i of the
87: *> matrix was interchanged with row IPIV(i).
88: *> \endverbatim
89: *>
90: *> \param[out] INFO
91: *> \verbatim
92: *> INFO is INTEGER
93: *> = 0: successful exit
94: *> < 0: if INFO = -i, the i-th argument had an illegal value
95: *> > 0: if INFO = i, U(i,i) is exactly zero. The factorization
96: *> has been completed, but the factor U is exactly
97: *> singular, and division by zero will occur if it is used
98: *> to solve a system of equations.
99: *> \endverbatim
100: *
101: * Authors:
102: * ========
103: *
1.4 bertrand 104: *> \author Univ. of Tennessee
105: *> \author Univ. of California Berkeley
106: *> \author Univ. of Colorado Denver
107: *> \author NAG Ltd.
1.1 bertrand 108: *
1.2 bertrand 109: *> \date June 2016
1.1 bertrand 110: *
111: *> \ingroup doubleGEcomputational
112: *
113: * =====================================================================
114: RECURSIVE SUBROUTINE DGETRF2( M, N, A, LDA, IPIV, INFO )
115: *
1.4 bertrand 116: * -- LAPACK computational routine (version 3.7.0) --
1.1 bertrand 117: * -- LAPACK is a software package provided by Univ. of Tennessee, --
118: * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
1.2 bertrand 119: * June 2016
1.1 bertrand 120: *
121: * .. Scalar Arguments ..
122: INTEGER INFO, LDA, M, N
123: * ..
124: * .. Array Arguments ..
125: INTEGER IPIV( * )
126: DOUBLE PRECISION A( LDA, * )
127: * ..
128: *
129: * =====================================================================
130: *
131: * .. Parameters ..
132: DOUBLE PRECISION ONE, ZERO
133: PARAMETER ( ONE = 1.0D+0, ZERO = 0.0D+0 )
134: * ..
135: * .. Local Scalars ..
136: DOUBLE PRECISION SFMIN, TEMP
137: INTEGER I, IINFO, N1, N2
138: * ..
139: * .. External Functions ..
140: DOUBLE PRECISION DLAMCH
141: INTEGER IDAMAX
142: EXTERNAL DLAMCH, IDAMAX
143: * ..
144: * .. External Subroutines ..
145: EXTERNAL DGEMM, DSCAL, DLASWP, DTRSM, XERBLA
146: * ..
147: * .. Intrinsic Functions ..
148: INTRINSIC MAX, MIN
149: * ..
150: * .. Executable Statements ..
151: *
152: * Test the input parameters
153: *
154: INFO = 0
155: IF( M.LT.0 ) THEN
156: INFO = -1
157: ELSE IF( N.LT.0 ) THEN
158: INFO = -2
159: ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
160: INFO = -4
161: END IF
162: IF( INFO.NE.0 ) THEN
163: CALL XERBLA( 'DGETRF2', -INFO )
164: RETURN
165: END IF
166: *
167: * Quick return if possible
168: *
169: IF( M.EQ.0 .OR. N.EQ.0 )
170: $ RETURN
171:
172: IF ( M.EQ.1 ) THEN
173: *
174: * Use unblocked code for one row case
175: * Just need to handle IPIV and INFO
176: *
177: IPIV( 1 ) = 1
178: IF ( A(1,1).EQ.ZERO )
179: $ INFO = 1
180: *
181: ELSE IF( N.EQ.1 ) THEN
182: *
183: * Use unblocked code for one column case
184: *
185: *
186: * Compute machine safe minimum
187: *
188: SFMIN = DLAMCH('S')
189: *
190: * Find pivot and test for singularity
191: *
192: I = IDAMAX( M, A( 1, 1 ), 1 )
193: IPIV( 1 ) = I
194: IF( A( I, 1 ).NE.ZERO ) THEN
195: *
196: * Apply the interchange
197: *
198: IF( I.NE.1 ) THEN
199: TEMP = A( 1, 1 )
200: A( 1, 1 ) = A( I, 1 )
201: A( I, 1 ) = TEMP
202: END IF
203: *
204: * Compute elements 2:M of the column
205: *
206: IF( ABS(A( 1, 1 )) .GE. SFMIN ) THEN
207: CALL DSCAL( M-1, ONE / A( 1, 1 ), A( 2, 1 ), 1 )
208: ELSE
209: DO 10 I = 1, M-1
210: A( 1+I, 1 ) = A( 1+I, 1 ) / A( 1, 1 )
211: 10 CONTINUE
212: END IF
213: *
214: ELSE
215: INFO = 1
216: END IF
217: *
218: ELSE
219: *
220: * Use recursive code
221: *
222: N1 = MIN( M, N ) / 2
223: N2 = N-N1
224: *
225: * [ A11 ]
226: * Factor [ --- ]
227: * [ A21 ]
228: *
229: CALL DGETRF2( M, N1, A, LDA, IPIV, IINFO )
230:
231: IF ( INFO.EQ.0 .AND. IINFO.GT.0 )
232: $ INFO = IINFO
233: *
234: * [ A12 ]
235: * Apply interchanges to [ --- ]
236: * [ A22 ]
237: *
238: CALL DLASWP( N2, A( 1, N1+1 ), LDA, 1, N1, IPIV, 1 )
239: *
240: * Solve A12
241: *
1.4 bertrand 242: CALL DTRSM( 'L', 'L', 'N', 'U', N1, N2, ONE, A, LDA,
1.1 bertrand 243: $ A( 1, N1+1 ), LDA )
244: *
245: * Update A22
246: *
1.4 bertrand 247: CALL DGEMM( 'N', 'N', M-N1, N2, N1, -ONE, A( N1+1, 1 ), LDA,
1.1 bertrand 248: $ A( 1, N1+1 ), LDA, ONE, A( N1+1, N1+1 ), LDA )
249: *
250: * Factor A22
251: *
252: CALL DGETRF2( M-N1, N2, A( N1+1, N1+1 ), LDA, IPIV( N1+1 ),
253: $ IINFO )
254: *
255: * Adjust INFO and the pivot indices
256: *
257: IF ( INFO.EQ.0 .AND. IINFO.GT.0 )
258: $ INFO = IINFO + N1
259: DO 20 I = N1+1, MIN( M, N )
260: IPIV( I ) = IPIV( I ) + N1
261: 20 CONTINUE
262: *
263: * Apply interchanges to A21
264: *
265: CALL DLASWP( N1, A( 1, 1 ), LDA, N1+1, MIN( M, N), IPIV, 1 )
266: *
267: END IF
268: RETURN
269: *
270: * End of DGETRF2
271: *
272: END
CVSweb interface <joel.bertrand@systella.fr>