Re: [deal.II] question about compute_projection_from_quadrature_points_matrix function

On 07/26/2016 06:21 AM, dealii.gr...@gmail.com wrote: I use a discontinuous field to transfer quadrature points data during mesh refinement. In addition I use same method to write quadrature data into output file. Why compute_projection_from_quadrature_points_matrix function sometime creates neg

Re: [deal.II] Problem using WorkStream with a hp::DoFHandler

On 07/26/2016 08:54 AM, Claire wrote: I hope that I provide enough information. Bruno suggestion was a good guess, but not the cause. The real cause is that your local assemble function takes the following arguments: const cell_iterator&, Assembly::ScratchData<3>&, Assembly::CopyData&

[deal.II] Re: How to solve each block separately when working with TrilinosWrappers::BlockSparseMatrix and AMG precondition.

Pasha, > *if* (direct_solver) > > { > > SolverControl cn; > > TrilinosWrappers::SolverDirect solver(cn); > > solver.*solve*(system_pde_matrix.block(0,0), > > newton_update.block(0), system_pde_residual.block(0)); > > > constraints_update.distribute(newton_update); > > > *return* 1; > > } > > > >

[deal.II] Re: Problem using WorkStream with a hp::DoFHandler

Claire, this looks to be the problem: /home/claire/Logiciels/deal.II/deal.II-v8.4.1/include/deal.II/base/work_stream.h:1230:3: note: template argument deduction/substitution failed: /home/claire/Logiciels/deal.II/deal.II-v8.4.1/Doctorat/slm_v18/slm_v18.cc:1031:17: note: deduced conflicting

[deal.II] Re: PETScWrappers::SparseDirectMUMPS for symmetry system matrix and factorization resue

Hi Alexander, I switch to : dealii 8.4.1 petsc 3.6.4 mumps 5.0.1 The result is still not correct. The modification I did for the code is replacing the following code in step-17.cc: *SolverControl solver_control (solution.size(),* *1e-8*s

[deal.II] Problem using WorkStream with a hp::DoFHandler

Dear all, First of all I would like to thank all the deal.II developers and the community who helped created it and keep it alive. The more I get into it the more I am impress by its high quality and completeness. The issue that I am currently facing is that I am trying to implement multithre

[deal.II] question about compute_projection_from_quadrature_points_matrix function

Dear all I use a discontinuous field to transfer quadrature points data during mesh refinement. In addition I use same method to write quadrature data into output file. Why compute_projection_from_quadrature_points_matrix function sometime creates negative data for a quadrature data which is a

Re: [deal.II] Re: PETScWrappers::SparseDirectMUMPS for symmetry system matrix and factorization resue

Hi Alexander, > On 26 Jul 2016, at 09:40, Alexander wrote: > > > @Denis: I agree, the logic is somewhat deficient. I also think documentation > and interface itself (e.g. I have no idea why does method set_solver_type >

[deal.II] Re: How to solve each block separately when working with TrilinosWrappers::BlockSparseMatrix and AMG precondition.

Dear Arndt Thank you for answer. I want to solve the system during a time step using a split method. First assembling block(0,0) and solving the displacement field, then using the updated displacement, the second field must be solved and so block(1,1) of system matrix is determined. I hav

[deal.II] Re: PETScWrappers::SparseDirectMUMPS for symmetry system matrix and factorization resue

> deal.II 8.4.0 > petsc 3.3.0.3 > MUMPS 4.10.0-p3 > This PETSc and MUMPS > 4 years old. Update to the latest versions and try again. If problem persists, provide the relevant snippet of the code. > Sorry, I am not familiar with the petsc library. You meant if I call > *solver.solve(system_m