CityOfHope

Hi @nmnobre , thanks a lot for the hints, I need bit more time to check it carefully, will come back as soon as possible.

Sorry for late reply. I still need some more time to go through the code, but the solution is exactly in https://github.com/mfem/mfem/issues/3209 pointed out by @nmnobre. Thanks again for your...

Hi, Will, thanks a lot for this great help, especially for a new MFEM user like me. I will give another try following the above example from you, will report...

HI Will. After setting up the system matrix. I tried to use preconditioned minres to solve the system: MINRESSolver solver(MPI_COMM_WORLD); solver.SetOperator(D); but D is a OperatorHandle object, it seems not...

Thanks Will. In the last step, to add the block diagonal preconditioner (diagonal of A, minus diagonal of C), I set the following: ```c++ // basically, both matrices A and...

Thanks a lot Will for your suggestions and comments. I will go through the code carefully, and will report soon if any problems pop up. I am very sorry that...

Hi Will. It seems the minres check the norm of the unpreconditioned residual, see, e.g., ex5p.cpp: // 13. Solve the linear system with MINRES. // Check the norm of the...

Hi all, thank you very much for clarification. Did I understand it correctly that the residual by MINRES is indeed in the preconditioner norm, as pointed out by you ?

Hi, Will. It seems still no convergence in the FEM discretization. Would it be possible to help check my rhs assembly (basically from exp5p.cpp), where `ufespace` and `pfespace` are standard...

Hi, a strange thing is even when I change fcoeff, it produce almost the same results, which is a sin(pi x) sin (pi y) sin (pi z) in 3D. Most...