Re: [deal.II] Re: How to shift a ConstraintMatrix that initialized with IndexSet?

Hi Daniel, In the source of ConstraintMatrix::shift, there is an assertion that the local_lines must be empty. Please see ConstraintMatrix::shift . Best, Ce Qin 2016-12-30 23:34 GMT+08:00 Daniel Arndt :

[deal.II] Re: Running Step-40 with eclipse PTP

Correction to the code above: mpi_initFinalize is not commented in the code. On Friday, December 30, 2016 at 6:05:03 PM UTC+1, Kartik Jujare wrote: > > Hello, > > I am trying to run the step 40 through eclipse PTP but have been > unsuccessful in running it on more than one processor. > > The foll

[deal.II] Running Step-40 with eclipse PTP

Hello, I am trying to run the step 40 through eclipse PTP but have been unsuccessful in running it on more than one processor. The following are the steps I tried: a) Step - 40 compiles and runs on two processors when run from the terminal. b) The template MPI Pi C++ project runs on two processo

[deal.II] Re: How to shift a ConstraintMatrix that initialized with IndexSet?

Ce Qin, Am Freitag, 30. Dezember 2016 02:46:05 UTC+1 schrieb Ce Qin: > > Dear all, > > I want to build a 2x2 block matrix, while all blocks are built from a > smaller DoFHandler. I need to generate 2 ConstraintMatrix and shift one > of them, then merge them together. In the source code, Timo Hei

Re: [deal.II] Dimension independent Maxwell`s equations

Thank you a lot! I`ve tried to play around and my conclusion is trivial and exactly what you have said - we do really need to solve a full 3d equation for 2D geometry too, it comes from physics, not from coding or math. Just imagine a typical 1D (or 2D) problem - an infinite subwavelength plate