1gesolver(1) Utility Commands gesolver(1)
2
3
4
6 gesolve - eigensolver for generalized eigenvalue problems (Ritz values
7 only)
8
9
11 gesolve matrix_a_filename matrix_b_filename evalues_filename [options]
12
13
15 This program inputs the matrix data from matrix_a_filename
16 matrix_b_filename, and solves the generalized eigenvalue problem A*x =
17 l*B*x with the solver specified by options. It outputs the specified
18 number of Ritz values, the number of which is given by option -ss, to
19 evalues_filename in the extended Matrix Market format (see Appendix of
20 Lis User Guide). Both the Matrix Market format and the Harwell-Boeing
21 format are supported for the matrix filenames.
22
23
25 The following options are supported:
26
27 -e eigensolver
28 The following options are supported for eigensolver:
29
30 -e {gli|15}
31 Generalized Lanczos
32
33 -ss [1]
34 The size of the subspace
35
36 -e {gai|16}
37 Generalized Arnoldi
38
39 -ss [1]
40 The size of the subspace
41
42 Other Options for eigensolver:
43
44 -eprint [0]
45 The output of the residual history
46
47 -eprint {none|0}
48 None
49
50 -eprint {mem|1}
51 Save the residual history
52
53 -eprint {out|2}
54 Output it to the standard output
55
56 -eprint {all|3}
57 Save the residual history and output it to the standard
58 output
59
60 -omp_num_threads [t]
61 The number of the threads (t represents the maximum number of
62 the threads)
63
64 -estorage [0]
65 The matrix storage format
66
67 -estorage_block [2]
68 The block size of the BSR and BSC formats
69
70 -ef [0]
71 The precision of the eigensolver
72
73 -ef {double|0}
74 Double precision
75
76 -ef {quad|1}
77 Double-double (quadruple) precision
78
79 See Lis User Guide for full description.
80
81
83 The following exit values are returned:
84
85 0 The process is normally terminated
86
87 unspecified
88 An error occurred
89
90
92 lis(3), lsolve(1), gesolve(1), hpcg_kernel(1), hpcg_spmvtest(1),
93 spmvtest1(1), spmvtest2(1), spmvtest2b(1), spmvtest3(1), spmvtest3b(1),
94 spmvtest4(1), spmvtest5(1)
95
96 http://www.ssisc.org/lis/
97 http://math.nist.gov/MatrixMarket/
98
99
100
101
102Man Page 4 Nov 2017 gesolver(1)