1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
|
C> \brief \b DGETRF VARIANT: iterative version of Sivan Toledo's recursive LU algorithm
*
* =========== DOCUMENTATION ===========
*
* Online html documentation available at
* http://www.netlib.org/lapack/explore-html/
*
* Definition:
* ===========
*
* SUBROUTINE DGETRF( M, N, A, LDA, IPIV, INFO )
*
* .. Scalar Arguments ..
* INTEGER INFO, LDA, M, N
* ..
* .. Array Arguments ..
* INTEGER IPIV( * )
* DOUBLE PRECISION A( LDA, * )
* ..
*
* Purpose
* =======
*
C>\details \b Purpose:
C>\verbatim
C>
C> DGETRF computes an LU factorization of a general M-by-N matrix A
C> using partial pivoting with row interchanges.
C>
C> The factorization has the form
C> A = P * L * U
C> where P is a permutation matrix, L is lower triangular with unit
C> diagonal elements (lower trapezoidal if m > n), and U is upper
C> triangular (upper trapezoidal if m < n).
C>
C> This code implements an iterative version of Sivan Toledo's recursive
C> LU algorithm[1]. For square matrices, this iterative versions should
C> be within a factor of two of the optimum number of memory transfers.
C>
C> The pattern is as follows, with the large blocks of U being updated
C> in one call to DTRSM, and the dotted lines denoting sections that
C> have had all pending permutations applied:
C>
C> 1 2 3 4 5 6 7 8
C> +-+-+---+-------+------
C> | |1| | |
C> |.+-+ 2 | |
C> | | | | |
C> |.|.+-+-+ 4 |
C> | | | |1| |
C> | | |.+-+ |
C> | | | | | |
C> |.|.|.|.+-+-+---+ 8
C> | | | | | |1| |
C> | | | | |.+-+ 2 |
C> | | | | | | | |
C> | | | | |.|.+-+-+
C> | | | | | | | |1|
C> | | | | | | |.+-+
C> | | | | | | | | |
C> |.|.|.|.|.|.|.|.+-----
C> | | | | | | | | |
C>
C> The 1-2-1-4-1-2-1-8-... pattern is the position of the last 1 bit in
C> the binary expansion of the current column. Each Schur update is
C> applied as soon as the necessary portion of U is available.
C>
C> [1] Toledo, S. 1997. Locality of Reference in LU Decomposition with
C> Partial Pivoting. SIAM J. Matrix Anal. Appl. 18, 4 (Oct. 1997),
C> 1065-1081. http://dx.doi.org/10.1137/S0895479896297744
C>
C>\endverbatim
*
* Arguments:
* ==========
*
C> \param[in] M
C> \verbatim
C> M is INTEGER
C> The number of rows of the matrix A. M >= 0.
C> \endverbatim
C>
C> \param[in] N
C> \verbatim
C> N is INTEGER
C> The number of columns of the matrix A. N >= 0.
C> \endverbatim
C>
C> \param[in,out] A
C> \verbatim
C> A is DOUBLE PRECISION array, dimension (LDA,N)
C> On entry, the M-by-N matrix to be factored.
C> On exit, the factors L and U from the factorization
C> A = P*L*U; the unit diagonal elements of L are not stored.
C> \endverbatim
C>
C> \param[in] LDA
C> \verbatim
C> LDA is INTEGER
C> The leading dimension of the array A. LDA >= max(1,M).
C> \endverbatim
C>
C> \param[out] IPIV
C> \verbatim
C> IPIV is INTEGER array, dimension (min(M,N))
C> The pivot indices; for 1 <= i <= min(M,N), row i of the
C> matrix was interchanged with row IPIV(i).
C> \endverbatim
C>
C> \param[out] INFO
C> \verbatim
C> INFO is INTEGER
C> = 0: successful exit
C> < 0: if INFO = -i, the i-th argument had an illegal value
C> > 0: if INFO = i, U(i,i) is exactly zero. The factorization
C> has been completed, but the factor U is exactly
C> singular, and division by zero will occur if it is used
C> to solve a system of equations.
C> \endverbatim
C>
*
* Authors:
* ========
*
C> \author Univ. of Tennessee
C> \author Univ. of California Berkeley
C> \author Univ. of Colorado Denver
C> \author NAG Ltd.
*
C> \date November 2011
*
C> \ingroup variantsGEcomputational
*
* =====================================================================
SUBROUTINE DGETRF( M, N, A, LDA, IPIV, INFO )
*
* -- LAPACK computational routine (version 3.X) --
* -- LAPACK is a software package provided by Univ. of Tennessee, --
* -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
* November 2011
*
* .. Scalar Arguments ..
INTEGER INFO, LDA, M, N
* ..
* .. Array Arguments ..
INTEGER IPIV( * )
DOUBLE PRECISION A( LDA, * )
* ..
*
* =====================================================================
*
* .. Parameters ..
DOUBLE PRECISION ONE, ZERO, NEGONE
PARAMETER ( ONE = 1.0D+0, ZERO = 0.0D+0 )
PARAMETER ( NEGONE = -1.0D+0 )
* ..
* .. Local Scalars ..
DOUBLE PRECISION SFMIN, TMP
INTEGER I, J, JP, NSTEP, NTOPIV, NPIVED, KAHEAD
INTEGER KSTART, IPIVSTART, JPIVSTART, KCOLS
* ..
* .. External Functions ..
DOUBLE PRECISION DLAMCH
INTEGER IDAMAX
LOGICAL DISNAN
EXTERNAL DLAMCH, IDAMAX, DISNAN
* ..
* .. External Subroutines ..
EXTERNAL DTRSM, DSCAL, XERBLA, DLASWP
* ..
* .. Intrinsic Functions ..
INTRINSIC MAX, MIN, IAND
* ..
* .. Executable Statements ..
*
* Test the input parameters.
*
INFO = 0
IF( M.LT.0 ) THEN
INFO = -1
ELSE IF( N.LT.0 ) THEN
INFO = -2
ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
INFO = -4
END IF
IF( INFO.NE.0 ) THEN
CALL XERBLA( 'DGETRF', -INFO )
RETURN
END IF
*
* Quick return if possible
*
IF( M.EQ.0 .OR. N.EQ.0 )
$ RETURN
*
* Compute machine safe minimum
*
SFMIN = DLAMCH( 'S' )
*
NSTEP = MIN( M, N )
DO J = 1, NSTEP
KAHEAD = IAND( J, -J )
KSTART = J + 1 - KAHEAD
KCOLS = MIN( KAHEAD, M-J )
*
* Find pivot.
*
JP = J - 1 + IDAMAX( M-J+1, A( J, J ), 1 )
IPIV( J ) = JP
* Permute just this column.
IF (JP .NE. J) THEN
TMP = A( J, J )
A( J, J ) = A( JP, J )
A( JP, J ) = TMP
END IF
* Apply pending permutations to L
NTOPIV = 1
IPIVSTART = J
JPIVSTART = J - NTOPIV
DO WHILE ( NTOPIV .LT. KAHEAD )
CALL DLASWP( NTOPIV, A( 1, JPIVSTART ), LDA, IPIVSTART, J,
$ IPIV, 1 )
IPIVSTART = IPIVSTART - NTOPIV;
NTOPIV = NTOPIV * 2;
JPIVSTART = JPIVSTART - NTOPIV;
END DO
* Permute U block to match L
CALL DLASWP( KCOLS, A( 1,J+1 ), LDA, KSTART, J, IPIV, 1 )
* Factor the current column
IF( A( J, J ).NE.ZERO .AND. .NOT.DISNAN( A( J, J ) ) ) THEN
IF( ABS(A( J, J )) .GE. SFMIN ) THEN
CALL DSCAL( M-J, ONE / A( J, J ), A( J+1, J ), 1 )
ELSE
DO I = 1, M-J
A( J+I, J ) = A( J+I, J ) / A( J, J )
END DO
END IF
ELSE IF( A( J,J ) .EQ. ZERO .AND. INFO .EQ. 0 ) THEN
INFO = J
END IF
* Solve for U block.
CALL DTRSM( 'Left', 'Lower', 'No transpose', 'Unit', KAHEAD,
$ KCOLS, ONE, A( KSTART, KSTART ), LDA,
$ A( KSTART, J+1 ), LDA )
* Schur complement.
CALL DGEMM( 'No transpose', 'No transpose', M-J,
$ KCOLS, KAHEAD, NEGONE, A( J+1, KSTART ), LDA,
$ A( KSTART, J+1 ), LDA, ONE, A( J+1, J+1 ), LDA )
END DO
* Handle pivot permutations on the way out of the recursion
NPIVED = IAND( NSTEP, -NSTEP )
J = NSTEP - NPIVED
DO WHILE ( J .GT. 0 )
NTOPIV = IAND( J, -J )
CALL DLASWP( NTOPIV, A( 1, J-NTOPIV+1 ), LDA, J+1, NSTEP,
$ IPIV, 1 )
J = J - NTOPIV
END DO
* If short and wide, handle the rest of the columns.
IF ( M .LT. N ) THEN
CALL DLASWP( N-M, A( 1, M+KCOLS+1 ), LDA, 1, M, IPIV, 1 )
CALL DTRSM( 'Left', 'Lower', 'No transpose', 'Unit', M,
$ N-M, ONE, A, LDA, A( 1,M+KCOLS+1 ), LDA )
END IF
RETURN
*
* End of DGETRF
*
END
|
