On Tue, 29 Nov 2016, Salazar De Troya, Miguel wrote:
On 11/17/16, 7:56 AM, "Roy Stogner" wrote:
I'd say:
Add a DiffSolver::linear_solver() API for returning a reference to a
LinearSolver, and add NewtonSolver and PetscDiffSolver
implementations.
For the PETSc case, that LinearSolver would
On 11/17/16, 7:56 AM, "Roy Stogner" wrote:
>I'd say:
>
>Add a DiffSolver::linear_solver() API for returning a reference to a
>LinearSolver, and add NewtonSolver and PetscDiffSolver
>implementations.
For the PETSc case, that LinearSolver would need to grab the KSP from the
SNES so we can reuse th
On Wed, 16 Nov 2016, Salazar De Troya, Miguel wrote:
LinearSolver does not have the capability to have null spaces attached to
it. In my opinion, wouldn’t it be a good idea to have a Solver class from
which all the solvers can inherit common things like attaching null spaces
or setting up the p
On Wed, 16 Nov 2016, Salazar De Troya, Miguel wrote:
The TimeSolver in DifferentiableSystem has both a DiffSolver and a
LinearSolver. The DiffSolver is used in the regular primal/forward solve()
and the LinearSolver in the adjoint_solve(). It would be convenient to
reuse the forward preconditio
LinearSolver does not have the capability to have null spaces attached to
it. In my opinion, wouldn’t it be a good idea to have a Solver class from
which all the solvers can inherit common things like attaching null spaces
or setting up the preconditioner?
On 11/16/16, 8:50 AM, "Salazar De Troya,
A slight tangent but still related to this topic of reusing the
preconditioner:
The TimeSolver in DifferentiableSystem has both a DiffSolver and a
LinearSolver. The DiffSolver is used in the regular primal/forward solve()
and the LinearSolver in the adjoint_solve(). It would be convenient to
reus
On Tue, 15 Nov 2016, Salazar De Troya, Miguel wrote:
What would be a good place to insert PetscDiffSolver::init()? Maybe
in DifferentiableSystem::init() after the user initialization?
I believe so.
---
Roy--
edu>>
Cc: Miguel Salazar mailto:salazardet...@llnl.gov>>,
"libmesh-users@lists.sourceforge.net<mailto:libmesh-users@lists.sourceforge.net>"
mailto:libmesh-users@lists.sourceforge.net>>
Subject: Re: [Libmesh-users] Recycling preconditioned in system_of_equations7.C
In petsc_nonlinear_solver, there is the same issue. After each solve, SNES
will get destroyed.Unfortunately, there are already a lot of
applications built on this wrong interface. However, we should fix this in
some point.
Fande,
On Tue, Nov 15, 2016 at 12:36 PM, Roy Stogner
wrote:
>
> On
On Tue, 15 Nov 2016, Salazar De Troya, Miguel wrote:
I am modifying PetscDiffSolve to be able to reuse the
preconditioner. However, every time PetscDiffSolver::solve() is
called, a new SNES is created by calling PetscDiffSolver::init(),
why is this?
I'm not sure, but if I had to bet money the
orge.net<mailto:libmesh-users@lists.sourceforge.net>"
mailto:libmesh-users@lists.sourceforge.net>>
Subject: Re: [Libmesh-users] Recycling preconditioned in system_of_equations7.C
On Fri, Nov 11, 2016 at 3:00 PM, Roy Stogner
mailto:royst...@ices.utexas.edu>> wrote:
On Th
On Fri, Nov 11, 2016 at 3:00 PM, Roy Stogner
wrote:
>
> On Thu, 10 Nov 2016, Salazar De Troya, Miguel wrote:
>
> > This example runs several linear solves because there is a nonlinear
> > solver below a continuation scheme. I figured that it would be
> > possible to reuse the preconditioner and s
On Thu, 10 Nov 2016, Salazar De Troya, Miguel wrote:
> This example runs several linear solves because there is a nonlinear
> solver below a continuation scheme. I figured that it would be
> possible to reuse the preconditioner and see savings so I included
> the lines:
>
> LinearSolver * linear_
Hi
This example runs several linear solves because there is a nonlinear solver
below a continuation scheme. I figured that it would be possible to reuse the
preconditioner and see savings so I included the lines:
LinearSolver * linear_solver = system.get_linear_solver();
linear_solver->reuse_
14 matches
Mail list logo
