1elpa_generalized_eigenvectorLsi(b3r)ary Functions Maenlupaal_generalized_eigenvectors(3)
2
3
4
6 elpa_generalized_eigenvectors - computes the generalized eigenvalues
7 and (part of) the eigenvector spectrum for a real symmetric or complex
8 hermitian matrix
9
10
12 FORTRAN INTERFACE
13 use elpa
14 class(elpa_t), pointer :: elpa
15
16 call elpa%generalized_eigenvectors (a, b, ev, q, is_already_decomopsed,
17 error)
18
19 With the definitions of the input and output variables:
20
21 class(elpa_t) :: elpa ! returns an instance of the ELPA object
22
23 datatype :: a
24 The matrix a for which the eigenvalues should be computed. The
25 dimensions of matrix a must be set BEFORE with the methods
26 elpa_set(3) and elpa_setup(3). The datatype of the matrix can be
27 one of "real(kind=c_double)", "real(kind=c_float)",
28 "complex(kind=c_double)", or "complex(kind=c_float)"
29
30 datatype :: b
31 The matrix b defining the generalized eigenvalue problem. The
32 dimensions and datatype of the matrix b has to be the same as
33 for matrix a.
34
35 datatype :: ev
36 The vector ev where the eigenvalues will be stored in ascending
37 order. The datatype of the vector ev can be either
38 "real(kind=c_double)", or "real(kind=c_float)", depending of the
39 datatype of the matrix. Note that complex hermitian matrices
40 also have real valued eigenvalues.
41
42 datatype :: q
43 The storage space for the computed eigenvectors. The dimensions
44 of matrix a must be set BEFORE with the methods elpa_set(3) and
45 elpa_setup(3). The datatype of the matrix can be one of
46 "real(kind=c_double)", "real(kind=c_float)",
47 "complex(kind=c_double)", or "complex(kind=c_float)"
48
49 logical :: is_already_decomposed
50 Has to be set to .false. for the first call with a given b and
51 .true. for each subsequent call with the same b, since b then
52 already contains decomposition and thus the decomposing step is
53 skipped.
54
55
56 integer, optional :: error
57 The return error code of the function. Should be "ELPA_OK". The
58 error code can be querried with the function elpa_strerr(3)
59
60
61 C INTERFACE
62 #include <elpa/elpa.h>
63 elpa_t handle;
64
65 void elpa_generalized_eigenvectors(elpa_t handle, datatype *a, datatype
66 *b, datatype *ev, datatype *q, int is_already_decomposed, int *error);
67
68 With the definitions of the input and output variables:
69
70
71 elpa_t handle;
72 The handle to the ELPA object
73
74 datatype *a;
75 The matrix a for which the eigenvalues should be computed. The
76 dimensions of the matrix must be set BEFORE with the methods
77 elpa_set(3) and elpa_setup(3). The datatype can be one of
78 "double", "float", "double complex", or "float complex".
79
80 datatype *b;
81 The matrix b defining the generalized eigenvalue problem. The
82 dimensions and the datatype of the matrix b must be the same as
83 matrix a.
84
85 datatype *ev;
86 The storage for the computed eigenvalues. Eigenvalues will be
87 stored in ascendig order. The datatype can be either "double" or
88 "float". Note that the eigenvalues of complex hermitian matrices
89 are also real.
90
91 datatype *q;
92 The storage space for the computed eigenvectors. The dimensions
93 of the matrix must be set BEFORE with the methods elpa_set(3)
94 and elpa_setup(3). The datatype can be one of "double", "float",
95 "double complex", or "float complex".
96
97 int is_already_decomposed;
98 Has to be set to 0 for the first call with a given b and 1 for
99 each subsequent call with the same b, since b then already
100 contains decomposition and thus the decomposing step is skipped.
101
102 int *error;
103 The error code of the function. Should be "ELPA_OK". The error
104 codes can be querried with elpa_strerr(3)
105
106
108 Compute the generalized eigenvalues and (parts of) the eigenvector
109 spectrum of a real symmtric or complex hermitian matrix.The functions
110 elpa_init(3), elpa_allocate(3), elpa_set(3), and elpa_setup(3) must be
111 called BEFORE elpa_generalized_eigenvalues can be called. Especially
112 the number of eigenvectors to be computed can be set with elpa_set(3).
113 Unlike in the case of ordinary eigenvalue problem, the generalized
114 problem calls some external scalapack routines. The user is responsible
115 for initialization of the blacs context, which then has to be passed to
116 elpa by elpa_set(3) BEFORE elpa_generalized_eigenvalues can be called.
117
119 elpa2_print_kernels(1) elpa_init(3) elpa_allocate(3) elpa_set(3)
120 elpa_setup(3) elpa_strerr(3) elpa_eigenvalues(3) elpa_eigenvectors(3)
121 elpa_cholesky(3) elpa_invert_triangular(3) elpa_solve_tridiagonal(3)
122 elpa_hermitian_multiply(3) elpa_uninit(3) elpa_deallocate(3)
123
124
125
126ELPA Thu Feb 1 2018elpa_generalized_eigenvectors(3)