1lsolve(1) Utility Commands lsolve(1)
2
6 lsolve - linear solver for sparse matrices
7
10 lsolve matrix_filename rhs_setting solution_filename rhistory_filename
11 [options]
12
15 This program inputs the data of the coefficient matrix from
16 matrix_filename and solves the linear equation A*x = b with the solver
17 specified by options. It outputs the solution to solution_filename in
18 the extended Matrix Market format and the residual history to rhis‐
19 tory_filename in the PLAIN format (see Appendix of the Lis User Guide).
20 Both the extended Matrix Market format and the Harwell-Boeing format
21 are supported for the matrix filename.
22 One of the following values can be specified by rhs_setting:
24 0 Use the right hand side vector b included in matrix_filename
26 1 Use b = (1, ..., 1)^T
28 2 Use b = A * (1, ..., 1)^T
30 rhs_filename
32 The filename for the right-hand side vector
33 The PLAIN and Matrix Market formats are supported for rhs_filename.
35
38 The following options are supported:
39 -i linear solver
41 The following options are supported for linear solver:
42 -i {cg|1}
44 CG
45 -i {bicg|2}
47 BiCG
48 -i {cgs|3}
50 CGS
51 -i {bicgstab|4}
53 BiCGSTAB
54 -i {bicgstabl|5}
56 BiCGSTAB(l)
57 -ell [2]
59 The degree l
60 -i {gpbicg|6}
62 GPBiCG
63 -i {tfqmr|7}
65 TFQMR
66 -i {orthomin|8}
68 Orthomin(m)
69 -restart [40]
71 The restart value m
72 -i {gmres|9}
74 GMRES(m)
75 -restart [40]
77 The restart value m
78 -i {jacobi|10}
80 Jacobi
81 -i {gs|11}
83 Gauss-Seidel
84 -i {sor|12}
86 SOR
87 -omega [1.9]
89 The relaxation coefficient omega (0<omega<2)
90 -i {bicgsafe|13}
92 BiCGSafe
93 -i {cr|14}
95 CR
96 -i {bicr|15}
98 BiCR
99 -i {crs|16}
101 CRS
102 -i {bicrstab|17}
104 BiCRSTAB
105 -i {gpbicr|18}
107 GPBiCR
108 -i {bicrsafe|19}
110 BiCRSafe
111 -i {fgmres|20}
113 FGMRES(m)
114 -restart [40]
116 The restart value m
117 -i {idrs|21}
119 IDR(s)
120 -irestart [2]
122 The restart value s
123 -i {idr1|22}
125 IDR(1)
126 -i {minres|23}
128 MINRES
129 -i {COCG|24}
131 COCG
132 -i {COCR|25}
134 COCR
135 -p preconditioner
138 The following options are supported for preconditioner:
139 -p {none|0}
141 None
142 -p {jacobi|1}
144 Jacobi
145 -p {ilu|2}
147 ILU(k)
148 -ilu_fill [0]
150 The fill level k
151 -p {ssor|3}
153 SSOR
154 -ssor_omega [1.0]
156 The relaxation coefficient omega (0<omega<2)
157 -p {hybrid|4}
159 Hybrid
160 -hybrid_i [sor]
162 The linear solver
163 -hybrid_maxiter [25]
165 The maximum number of the iterations
166 -hybrid_tol [1.0e-3]
168 The convergence criterion
169 -hybrid_omega [1.5]
171 The relaxation coefficient omega of the SOR
172 (0<omega<2)
173 -hybrid_ell [2]
175 The degree l of the BiCGSTAB(l)
176 -hybrid_restart [40]
178 The restart values of the GMRES and Orthomin
179 -p {is|5}
181 I+S
182 -is_alpha [1.0]
184 The parameter alpha of I+alpha*S(m)
185 -is_m [3]
187 The parameter m of I+alpha*S(m)
188 -p {sainv|6}
190 SAINV
191 -sainv_drop [0.05]
193 The drop criterion
194 -p {saamg|7}
196 SA-AMG
197 -saamg_unsym [false]
199 Select the unsymmetric version (The matrix struc‐
200 ture must be symmetric)
201 -saamg_theta [0.05|0.12]
203 The drop criterion
204 -p {iluc|8}
206 Crout ILU
207 -iluc_drop [0.05]
209 The drop criterion
210 -iluc_rate [5.0]
212 The ration of maximum fill-in
213 -p {ilut|9}
215 ILUT
216 -ilut_drop [0.05]
218 The drop criterion
219 -ilut_rate [5.0]
221 The ration of maximum fill-in
222 -adds true
224 Additive Schwarz
225 -adds_iter [1]
227 The number of the iteration
228 Other Options:
230 -maxiter [1000]
232 The maximum number of the iterations
233 -tol [1.0e-12]
235 The convergence criterion
236 -print [0]
238 The output of the residual history
239 -print {none|0}
241 None
242 -print {mem|1}
244 Save the residual history
245 -print {out|2}
247 Output it to the standard output
248 -print {all|3}
250 Save the residual history and output it to the standard
251 output
252 -scale [0]
254 The scaling
255 -scale {none|0}
257 No scaling
258 -scale {jacobi|1}
260 The Jacobi scaling
261 -scale {symm_diag|2}
263 The diagonal scaling
264 -initx_zeros [true]
266 The behavior of the initial vector x_0
267 -initx_zero {false|0}
269 Given values
270 -initx_zero {true|1}
272 All values are set to 0
273 -omp_num_threads [t]
275 The number of the threads (t represents the maximum number of
276 the threads)
277 -storage [0]
279 The matrix storage format
280 -storage_block [2]
282 The block size of the BSR and BSC formats
283 -f [0] The precision of the linear solver
285 -f {double|0}
287 Double precision
288 -f {quad|1}
290 Double-double (quadruple) precision
291 See Lis User Guide for full description.
293
296 The following exit values are returned:
297 0 The process is normally terminated
299 unspecified
301 An error occurred
302
305 lis(3), esolve(1), hpcg_kernel(1), hpcg_spmvtest(1), spmvtest1(1),
306 spmvtest2(1), spmvtest2b(1), spmvtest3(1), spmvtest3b(1), spmvtest4(1),
307 spmvtest5(1)
308 http://www.ssisc.org/lis/
310 http://math.nist.gov/MatrixMarket/
311Man Page 14 Sep 2017 lsolve(1)