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
|
---
:name: zgeevx
:md5sum: c5b0634690cc9242af9ab759a937a099
:category: :subroutine
:arguments:
- balanc:
:type: char
:intent: input
- jobvl:
:type: char
:intent: input
- jobvr:
:type: char
:intent: input
- sense:
:type: char
:intent: input
- n:
:type: integer
:intent: input
- a:
:type: doublecomplex
:intent: input/output
:dims:
- lda
- n
- lda:
:type: integer
:intent: input
- w:
:type: doublecomplex
:intent: output
:dims:
- n
- vl:
:type: doublecomplex
:intent: output
:dims:
- ldvl
- n
- ldvl:
:type: integer
:intent: input
- vr:
:type: doublecomplex
:intent: output
:dims:
- ldvr
- n
- ldvr:
:type: integer
:intent: input
- ilo:
:type: integer
:intent: output
- ihi:
:type: integer
:intent: output
- scale:
:type: doublereal
:intent: output
:dims:
- n
- abnrm:
:type: doublereal
:intent: output
- rconde:
:type: doublereal
:intent: output
:dims:
- n
- rcondv:
:type: doublereal
:intent: output
:dims:
- n
- work:
:type: doublecomplex
:intent: output
:dims:
- MAX(1,lwork)
- lwork:
:type: integer
:intent: input
:option: true
:default: "(lsame_(&sense,\"N\")||lsame_(&sense,\"E\")) ? 2*n : (lsame_(&sense,\"V\")||lsame_(&sense,\"B\")) ? n*n+2*n : 0"
- rwork:
:type: doublereal
:intent: workspace
:dims:
- 2*n
- info:
:type: integer
:intent: output
:substitutions:
ldvr: "lsame_(&jobvr,\"V\") ? n : 1"
ldvl: "lsame_(&jobvl,\"V\") ? n : 1"
:fortran_help: " SUBROUTINE ZGEEVX( BALANC, JOBVL, JOBVR, SENSE, N, A, LDA, W, VL, LDVL, VR, LDVR, ILO, IHI, SCALE, ABNRM, RCONDE, RCONDV, WORK, LWORK, RWORK, INFO )\n\n\
* Purpose\n\
* =======\n\
*\n\
* ZGEEVX computes for an N-by-N complex nonsymmetric matrix A, the\n\
* eigenvalues and, optionally, the left and/or right eigenvectors.\n\
*\n\
* Optionally also, it computes a balancing transformation to improve\n\
* the conditioning of the eigenvalues and eigenvectors (ILO, IHI,\n\
* SCALE, and ABNRM), reciprocal condition numbers for the eigenvalues\n\
* (RCONDE), and reciprocal condition numbers for the right\n\
* eigenvectors (RCONDV).\n\
*\n\
* The right eigenvector v(j) of A satisfies\n\
* A * v(j) = lambda(j) * v(j)\n\
* where lambda(j) is its eigenvalue.\n\
* The left eigenvector u(j) of A satisfies\n\
* u(j)**H * A = lambda(j) * u(j)**H\n\
* where u(j)**H denotes the conjugate transpose of u(j).\n\
*\n\
* The computed eigenvectors are normalized to have Euclidean norm\n\
* equal to 1 and largest component real.\n\
*\n\
* Balancing a matrix means permuting the rows and columns to make it\n\
* more nearly upper triangular, and applying a diagonal similarity\n\
* transformation D * A * D**(-1), where D is a diagonal matrix, to\n\
* make its rows and columns closer in norm and the condition numbers\n\
* of its eigenvalues and eigenvectors smaller. The computed\n\
* reciprocal condition numbers correspond to the balanced matrix.\n\
* Permuting rows and columns will not change the condition numbers\n\
* (in exact arithmetic) but diagonal scaling will. For further\n\
* explanation of balancing, see section 4.10.2 of the LAPACK\n\
* Users' Guide.\n\
*\n\n\
* Arguments\n\
* =========\n\
*\n\
* BALANC (input) CHARACTER*1\n\
* Indicates how the input matrix should be diagonally scaled\n\
* and/or permuted to improve the conditioning of its\n\
* eigenvalues.\n\
* = 'N': Do not diagonally scale or permute;\n\
* = 'P': Perform permutations to make the matrix more nearly\n\
* upper triangular. Do not diagonally scale;\n\
* = 'S': Diagonally scale the matrix, ie. replace A by\n\
* D*A*D**(-1), where D is a diagonal matrix chosen\n\
* to make the rows and columns of A more equal in\n\
* norm. Do not permute;\n\
* = 'B': Both diagonally scale and permute A.\n\
*\n\
* Computed reciprocal condition numbers will be for the matrix\n\
* after balancing and/or permuting. Permuting does not change\n\
* condition numbers (in exact arithmetic), but balancing does.\n\
*\n\
* JOBVL (input) CHARACTER*1\n\
* = 'N': left eigenvectors of A are not computed;\n\
* = 'V': left eigenvectors of A are computed.\n\
* If SENSE = 'E' or 'B', JOBVL must = 'V'.\n\
*\n\
* JOBVR (input) CHARACTER*1\n\
* = 'N': right eigenvectors of A are not computed;\n\
* = 'V': right eigenvectors of A are computed.\n\
* If SENSE = 'E' or 'B', JOBVR must = 'V'.\n\
*\n\
* SENSE (input) CHARACTER*1\n\
* Determines which reciprocal condition numbers are computed.\n\
* = 'N': None are computed;\n\
* = 'E': Computed for eigenvalues only;\n\
* = 'V': Computed for right eigenvectors only;\n\
* = 'B': Computed for eigenvalues and right eigenvectors.\n\
*\n\
* If SENSE = 'E' or 'B', both left and right eigenvectors\n\
* must also be computed (JOBVL = 'V' and JOBVR = 'V').\n\
*\n\
* N (input) INTEGER\n\
* The order of the matrix A. N >= 0.\n\
*\n\
* A (input/output) COMPLEX*16 array, dimension (LDA,N)\n\
* On entry, the N-by-N matrix A.\n\
* On exit, A has been overwritten. If JOBVL = 'V' or\n\
* JOBVR = 'V', A contains the Schur form of the balanced\n\
* version of the matrix A.\n\
*\n\
* LDA (input) INTEGER\n\
* The leading dimension of the array A. LDA >= max(1,N).\n\
*\n\
* W (output) COMPLEX*16 array, dimension (N)\n\
* W contains the computed eigenvalues.\n\
*\n\
* VL (output) COMPLEX*16 array, dimension (LDVL,N)\n\
* If JOBVL = 'V', the left eigenvectors u(j) are stored one\n\
* after another in the columns of VL, in the same order\n\
* as their eigenvalues.\n\
* If JOBVL = 'N', VL is not referenced.\n\
* u(j) = VL(:,j), the j-th column of VL.\n\
*\n\
* LDVL (input) INTEGER\n\
* The leading dimension of the array VL. LDVL >= 1; if\n\
* JOBVL = 'V', LDVL >= N.\n\
*\n\
* VR (output) COMPLEX*16 array, dimension (LDVR,N)\n\
* If JOBVR = 'V', the right eigenvectors v(j) are stored one\n\
* after another in the columns of VR, in the same order\n\
* as their eigenvalues.\n\
* If JOBVR = 'N', VR is not referenced.\n\
* v(j) = VR(:,j), the j-th column of VR.\n\
*\n\
* LDVR (input) INTEGER\n\
* The leading dimension of the array VR. LDVR >= 1; if\n\
* JOBVR = 'V', LDVR >= N.\n\
*\n\
* ILO (output) INTEGER\n\
* IHI (output) INTEGER\n\
* ILO and IHI are integer values determined when A was\n\
* balanced. The balanced A(i,j) = 0 if I > J and\n\
* J = 1,...,ILO-1 or I = IHI+1,...,N.\n\
*\n\
* SCALE (output) DOUBLE PRECISION array, dimension (N)\n\
* Details of the permutations and scaling factors applied\n\
* when balancing A. If P(j) is the index of the row and column\n\
* interchanged with row and column j, and D(j) is the scaling\n\
* factor applied to row and column j, then\n\
* SCALE(J) = P(J), for J = 1,...,ILO-1\n\
* = D(J), for J = ILO,...,IHI\n\
* = P(J) for J = IHI+1,...,N.\n\
* The order in which the interchanges are made is N to IHI+1,\n\
* then 1 to ILO-1.\n\
*\n\
* ABNRM (output) DOUBLE PRECISION\n\
* The one-norm of the balanced matrix (the maximum\n\
* of the sum of absolute values of elements of any column).\n\
*\n\
* RCONDE (output) DOUBLE PRECISION array, dimension (N)\n\
* RCONDE(j) is the reciprocal condition number of the j-th\n\
* eigenvalue.\n\
*\n\
* RCONDV (output) DOUBLE PRECISION array, dimension (N)\n\
* RCONDV(j) is the reciprocal condition number of the j-th\n\
* right eigenvector.\n\
*\n\
* WORK (workspace/output) COMPLEX*16 array, dimension (MAX(1,LWORK))\n\
* On exit, if INFO = 0, WORK(1) returns the optimal LWORK.\n\
*\n\
* LWORK (input) INTEGER\n\
* The dimension of the array WORK. If SENSE = 'N' or 'E',\n\
* LWORK >= max(1,2*N), and if SENSE = 'V' or 'B',\n\
* LWORK >= N*N+2*N.\n\
* For good performance, LWORK must generally be larger.\n\
*\n\
* If LWORK = -1, then a workspace query is assumed; the routine\n\
* only calculates the optimal size of the WORK array, returns\n\
* this value as the first entry of the WORK array, and no error\n\
* message related to LWORK is issued by XERBLA.\n\
*\n\
* RWORK (workspace) DOUBLE PRECISION array, dimension (2*N)\n\
*\n\
* INFO (output) INTEGER\n\
* = 0: successful exit\n\
* < 0: if INFO = -i, the i-th argument had an illegal value.\n\
* > 0: if INFO = i, the QR algorithm failed to compute all the\n\
* eigenvalues, and no eigenvectors or condition numbers\n\
* have been computed; elements 1:ILO-1 and i+1:N of W\n\
* contain eigenvalues which have converged.\n\
*\n\n\
* =====================================================================\n\
*\n"
|
