Re: [deal.II] Linear operators - TrilinosWrappers

Dear Jean-Paul, Finally it is working! Thank you for your help and clear explanation! Now it is a bit more clear for me how these linear operators work here :). (Regarding the op_M operator, I was not using that variable because I declared M directly as the linear operator of the block(0,0).

Re: [deal.II] Linear operators - TrilinosWrappers

Dear Felipe, Changing this const auto op_pre = linear_operator(prec_M); to this const auto op_pre = linear_operator(M, prec_M); allowed your program "FEMDG-Dar-par_tofix" to compile for me. If you have a look at this documentation for this variant of linear_operator

Re: [deal.II] Linear operators - TrilinosWrappers

Dear Jean-Paul, Thank you for your reply, I already set up the Identity preconditioner with a block matrix, but the problem remains the same. I am not sure this is the problem because I didn't need to set up the identity operator to any matrix in the scenario where I could invert. I think my

Re: [deal.II] Linear operators - TrilinosWrappers

Dear Felipe, It might be that you need to set up the preconditioner operator with an exemplar matrix (since the identity operator doesn't know the size of the range and domain that it's working on). If that's not the issue then could you please try to reproduce this as a minimal example and

Re: [deal.II] Linear operators - TrilinosWrappers

Dear Jean-Paul, Thank you for your reply and the complete information you provide me. Effectively, declaring the preconditioner (out-of-line) was one problem, but the inverse_operator function is still not matching. What I just realized is that when I compute the operator as a multiplication

Re: [deal.II] Linear operators - TrilinosWrappers

Hi again Feilpe, Regarding the lack of documentation, I’ve opened an issue on Github to track this. You can find that here: https://github.com/dealii/dealii/issues/12466 Best, Jean-Paul > On 15. Jun 2021, at 12:57, Jean-Paul Pelteret wrote: >

Re: [deal.II] Linear operators - TrilinosWrappers

Hi Feilpe, Firstly, I agree that the documentation is very light on details on how to use the linear operators with Trilinos linear algebra types. We can definitely improve this, and any contributions to that effect would be greatly appreciated! Right now, I can direct you to a few of the

[deal.II] Linear operators - TrilinosWrappers

Hello everyone, I have implemented a residual-minimization framework that somehow is similar to DPG. I want to extend my results by using parallelization using MPI with PETSc or Trilinos. So far, I have solved the saddle point problem using the Schur complement exactly how it is described in