Lines Matching refs:be

52 Other defect-correction algorithms can be implemented by using different
62 can be approximated with finite differences. For codes that are
117 where `x` should be initialized to the initial guess before calling and contains the solution on re…
141 evaluation routine; `NULL` should be used if such information is not
143 which various objects can be stashed; in Fortran a user context can be
174 formats (including “matrix-free” methods) can be used.
192 `-snes_mf_operator` may be used to obtain the exact solution, where
209 would be linearized as
218 Picard iterative procedure can be performed by defining $\mathbf{J}(\mathbf{x})$
219 to be $\mathbf{A}(\mathbf{x})$. Sometimes this iteration exhibits better global
227 entirely new matrix context may be preferable. Upon subsequent calls to
247 This can be due to bugs in the application code or because the proposed solution is not in the doma…
251 If they are unable to recover the `SNESConvergedReason` returned will be `SNES_DIVERGED_FUNCTION_DO…
264 Each solver may have associated with it a set of options, which can be
266 purpose. A complete list can be found by consulting the manual pages or
385 may be retrieved using
392 polynomial approximation may be set with `-snes_linesearch_order` or
398 for instance, 2 for quadratic or 3 for cubic. Sometimes, it may not be
406 may be used to turn off function, step, and solution norm computation at
422 line searches, which may be used to find a better-fitted steplength in
424 iterations may be set with `-snes_linesearch_max_it`. In addition, the
425 convergence criteria of the iterative line searches may be set using
459 Custom line search types may either be defined using
471 MINPACK project {cite}`more84`. Several parameters can be
494 which can be thought of as applying the external force in proportional load
499 function, which allows for the current value of $\lambda$ to be queried in the
524 $\mathbf x$ and $\lambda$, we define the linearized equilibrium line to be
546 While a similar process could be implemented as a `TS`, this method is
547 particularly designed to be used as a SNES, either standalone or within a `TS`.
555 As such, the final increment will likely be a hybrid of arc length continuation and
647 are implemented in PETSc and may be chosen using
659 `SNES` object that may be obtained with `SNESGetNPC()`.
667 `-snes_type qn` and may be changed with `snes_qn_type`. The default
671 …use of a nonlinear preconditioner that can be obtained with `SNESGetNPC()` and then configured; or…
717 map. We describe the “get” operations here, and it can be assumed that
719 levels in the hierarchy may be retrieved using
727 type may be set with
733 and the cycle type, 1 for V, 2 for W, may be set with
740 various levels’ cycles and smoothers. The level smoothers may be
756 be acquired with
764 variables, called injection. This may be set with
770 The coarse solve context may be acquired with
791 nonoverlapping fashion. This may be set with
811 Convergence of the nonlinear solvers can be detected in a variety of
816 are set by default to values that should be reasonable for a wide range
822 than some tolerance, `stol`. Convergence can also be determined based
842 A related routine is `SNESGetTolerances()`. `PETSC_CURRENT` may be used
843 for any parameter to indicate the current value should be retained; use `PETSC_DETERMINE` to restor…
856 respectively. `SNESConvergedReason` should be set positive for
880 should be done after the solution update. The option `-snes_monitor`
914 Since hand-coding routines for Jacobian matrix evaluation can be error
921 causes the two matrices to be printed to the screen.
932 The correctness of user provided `MATSHELL` Jacobians in general can be
936 can be checked with `TSRHSJacobianTestTranspose()` and `TSRHSJacobianTest()`
938 matrix-product. From the command line, these can be checked with
948 at each iteration can be costly, modifications
970 techniques. While the Newton systems must be solved well enough to
977 A sequence of nonnegative numbers $\{\eta_k\}$ can be used to
1010 specified in the Jacobian evaluation routine need not be conventional
1017 is, obviously matrix-free methods cannot be used with a direct solver,
1032 matrix-free approximations can also be invoked with the command
1043 which can also be used directly for users who need a matrix-free matrix but are not using `SNES`.
1052 The parameter `rerror` should be set to the square root of the
1056 floating-point precision accuracy. This parameter can also be set from
1099 hence $\sqrt{1 + ||u||}$ need be computed only once. This
1100 information may be set with the options
1124 `SNESSetFromOptions()` must be called *after* `SNESSetJacobian()` to
1164 noise of the functions. This can be selected with `-snes_mf_version 2`; it does not use `MatCreateM…
1180 - Next, one creates a `MatFDColoring` data structure that will be
1202   rows. These columns in the Jacobian can all be computed simultaneously.
1227 structure is so that it may be colored? Determining the structure is
1233 …uctured meshes, the nonzero locations will be identified in `DMCreateMatrix()` and the procedure a…
1237 coloring. For example, if a grid can be colored directly (without using
1238 the associated matrix), then that coloring can be provided to
1243 PETSc provides the following coloring algorithms, which can be selected using `MatColoringSetType()…
1318 for `MATMFFD_WP` (default). These parameters may be set from the options
1353 $L_i \le u_i \le H_i$. For example, the pressure variable cannot be negative.
1354 Some, or all, of the lower bounds may be
1356 all, of the upper bounds may be infinity (indicated by `SNES_VI_INF`).
1364 …cate that one is solving a variational inequality. Problems with box constraints can be solved with
1372 `SNESLineSearchSetPreCheck()` and `SNESLineSearchSetPostCheck()` can also be used to control proper…
1382 instance may be extracted using
1389 preconditioning comes in two flavors: left and right. The side may be
1405 which can be interpreted as putting the preconditioner into “striking
1410 instances to be combined additively or multiplicatively. By command
1411 line, a set of `SNES` types may be given by comma separated list
1415 (`SNES_COMPOSITE_MULTIPLICATIVE`) variants which may be set with
1421 New subsolvers may be added to the composite solver with