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
|
---
:name: dlaed8
:md5sum: d8ce12a706721da54812694fab3a1bbe
:category: :subroutine
:arguments:
- icompq:
:type: integer
:intent: input
- k:
:type: integer
:intent: output
- n:
:type: integer
:intent: input
- qsiz:
:type: integer
:intent: input
- d:
:type: doublereal
:intent: input/output
:dims:
- n
- q:
:type: doublereal
:intent: input/output
:dims:
- "icompq==0 ? 0 : ldq"
- "icompq==0 ? 0 : n"
- ldq:
:type: integer
:intent: input
- indxq:
:type: integer
:intent: input
:dims:
- n
- rho:
:type: doublereal
:intent: input/output
- cutpnt:
:type: integer
:intent: input
- z:
:type: doublereal
:intent: input
:dims:
- n
- dlamda:
:type: doublereal
:intent: output
:dims:
- n
- q2:
:type: doublereal
:intent: output
:dims:
- "icompq==0 ? 0 : ldq2"
- "icompq==0 ? 0 : n"
- ldq2:
:type: integer
:intent: input
- w:
:type: doublereal
:intent: output
:dims:
- n
- perm:
:type: integer
:intent: output
:dims:
- n
- givptr:
:type: integer
:intent: output
- givcol:
:type: integer
:intent: output
:dims:
- "2"
- n
- givnum:
:type: doublereal
:intent: output
:dims:
- "2"
- n
- indxp:
:type: integer
:intent: workspace
:dims:
- n
- indx:
:type: integer
:intent: workspace
:dims:
- n
- info:
:type: integer
:intent: output
:substitutions:
ldq2: MAX(1,n)
:fortran_help: " SUBROUTINE DLAED8( ICOMPQ, K, N, QSIZ, D, Q, LDQ, INDXQ, RHO, CUTPNT, Z, DLAMDA, Q2, LDQ2, W, PERM, GIVPTR, GIVCOL, GIVNUM, INDXP, INDX, INFO )\n\n\
* Purpose\n\
* =======\n\
*\n\
* DLAED8 merges the two sets of eigenvalues together into a single\n\
* sorted set. Then it tries to deflate the size of the problem.\n\
* There are two ways in which deflation can occur: when two or more\n\
* eigenvalues are close together or if there is a tiny element in the\n\
* Z vector. For each such occurrence the order of the related secular\n\
* equation problem is reduced by one.\n\
*\n\n\
* Arguments\n\
* =========\n\
*\n\
* ICOMPQ (input) INTEGER\n\
* = 0: Compute eigenvalues only.\n\
* = 1: Compute eigenvectors of original dense symmetric matrix\n\
* also. On entry, Q contains the orthogonal matrix used\n\
* to reduce the original matrix to tridiagonal form.\n\
*\n\
* K (output) INTEGER\n\
* The number of non-deflated eigenvalues, and the order of the\n\
* related secular equation.\n\
*\n\
* N (input) INTEGER\n\
* The dimension of the symmetric tridiagonal matrix. N >= 0.\n\
*\n\
* QSIZ (input) INTEGER\n\
* The dimension of the orthogonal matrix used to reduce\n\
* the full matrix to tridiagonal form. QSIZ >= N if ICOMPQ = 1.\n\
*\n\
* D (input/output) DOUBLE PRECISION array, dimension (N)\n\
* On entry, the eigenvalues of the two submatrices to be\n\
* combined. On exit, the trailing (N-K) updated eigenvalues\n\
* (those which were deflated) sorted into increasing order.\n\
*\n\
* Q (input/output) DOUBLE PRECISION array, dimension (LDQ,N)\n\
* If ICOMPQ = 0, Q is not referenced. Otherwise,\n\
* on entry, Q contains the eigenvectors of the partially solved\n\
* system which has been previously updated in matrix\n\
* multiplies with other partially solved eigensystems.\n\
* On exit, Q contains the trailing (N-K) updated eigenvectors\n\
* (those which were deflated) in its last N-K columns.\n\
*\n\
* LDQ (input) INTEGER\n\
* The leading dimension of the array Q. LDQ >= max(1,N).\n\
*\n\
* INDXQ (input) INTEGER array, dimension (N)\n\
* The permutation which separately sorts the two sub-problems\n\
* in D into ascending order. Note that elements in the second\n\
* half of this permutation must first have CUTPNT added to\n\
* their values in order to be accurate.\n\
*\n\
* RHO (input/output) DOUBLE PRECISION\n\
* On entry, the off-diagonal element associated with the rank-1\n\
* cut which originally split the two submatrices which are now\n\
* being recombined.\n\
* On exit, RHO has been modified to the value required by\n\
* DLAED3.\n\
*\n\
* CUTPNT (input) INTEGER\n\
* The location of the last eigenvalue in the leading\n\
* sub-matrix. min(1,N) <= CUTPNT <= N.\n\
*\n\
* Z (input) DOUBLE PRECISION array, dimension (N)\n\
* On entry, Z contains the updating vector (the last row of\n\
* the first sub-eigenvector matrix and the first row of the\n\
* second sub-eigenvector matrix).\n\
* On exit, the contents of Z are destroyed by the updating\n\
* process.\n\
*\n\
* DLAMDA (output) DOUBLE PRECISION array, dimension (N)\n\
* A copy of the first K eigenvalues which will be used by\n\
* DLAED3 to form the secular equation.\n\
*\n\
* Q2 (output) DOUBLE PRECISION array, dimension (LDQ2,N)\n\
* If ICOMPQ = 0, Q2 is not referenced. Otherwise,\n\
* a copy of the first K eigenvectors which will be used by\n\
* DLAED7 in a matrix multiply (DGEMM) to update the new\n\
* eigenvectors.\n\
*\n\
* LDQ2 (input) INTEGER\n\
* The leading dimension of the array Q2. LDQ2 >= max(1,N).\n\
*\n\
* W (output) DOUBLE PRECISION array, dimension (N)\n\
* The first k values of the final deflation-altered z-vector and\n\
* will be passed to DLAED3.\n\
*\n\
* PERM (output) INTEGER array, dimension (N)\n\
* The permutations (from deflation and sorting) to be applied\n\
* to each eigenblock.\n\
*\n\
* GIVPTR (output) INTEGER\n\
* The number of Givens rotations which took place in this\n\
* subproblem.\n\
*\n\
* GIVCOL (output) INTEGER array, dimension (2, N)\n\
* Each pair of numbers indicates a pair of columns to take place\n\
* in a Givens rotation.\n\
*\n\
* GIVNUM (output) DOUBLE PRECISION array, dimension (2, N)\n\
* Each number indicates the S value to be used in the\n\
* corresponding Givens rotation.\n\
*\n\
* INDXP (workspace) INTEGER array, dimension (N)\n\
* The permutation used to place deflated values of D at the end\n\
* of the array. INDXP(1:K) points to the nondeflated D-values\n\
* and INDXP(K+1:N) points to the deflated eigenvalues.\n\
*\n\
* INDX (workspace) INTEGER array, dimension (N)\n\
* The permutation used to sort the contents of D into ascending\n\
* order.\n\
*\n\
* INFO (output) INTEGER\n\
* = 0: successful exit.\n\
* < 0: if INFO = -i, the i-th argument had an illegal value.\n\
*\n\n\
* Further Details\n\
* ===============\n\
*\n\
* Based on contributions by\n\
* Jeff Rutter, Computer Science Division, University of California\n\
* at Berkeley, USA\n\
*\n\
* =====================================================================\n\
*\n"
|
