[petsc-users] How to view draw a matrix?

[Florian Lindner]

Hey,

Am 04.11.2016 um 15:21 schrieb Matthew Knepley:
> On Fri, Nov 4, 2016 at 9:14 AM, Florian Lindner <mailinglists at xgm.de <mailto:mailinglists at xgm.de>> wrote:
> 
>     Hello,
> 
>     I want to print the sparsity structure of a matrix of type MPISBAIJ and try to use this code:
> 
>       PetscErrorCode ierr = 0;
>       PetscViewer viewer;
>       PetscDraw draw;
>       ierr = PetscViewerCreate(communicator, &viewer); CHKERRV(ierr);
>       ierr = PetscViewerSetType(viewer, PETSCVIEWERDRAW); CHKERRV(ierr);
>       ierr = MatView(matrix, viewer); CHKERRV(ierr);
>       ierr = PetscViewerDrawGetDraw(viewer, 0, &draw); CHKERRV(ierr);
>       ierr = PetscDrawSetPause(draw, -1); CHKERRV(ierr); // Wait for user
>       ierr = PetscDrawDestroy(&draw); CHKERRV(ierr);
>       ierr = PetscViewerDestroy(&viewer); CHKERRV(ierr);
> 
>     communicator is of type MPI_Comm and is PETSC_COMM_WORLD.
> 
>     It opens an X window and shows the matrix just fine. The window closes after a right click and then my application
>     crashes:
> 
>     --------------------------------------------------------------------------
>     [1]PETSC ERROR: PetscDrawFlush() line 27 in
>     /data/scratch/lindnefn/software/petsc/src/sys/classes/draw/interface/dflush.c
>         Wrong type of object: Parameter # 1
>     --------------------------------------------------------------------------
> 
>     I tried to swap the DrawDestroy and ViewerDestroy calls (though, imho the order is correct), but changes nothing.
> 
> 
> 1) Do not Destroy the Draw you get. That is just a borrowed reference.

Perfekt, that works.

> 2) You can do all that by calling
> 
>   MatViewFromOptions(matrix, NULL, "-mymat_view");
> 
> and then passing
> 
>   -mymat_view draw -draw_pause -1

Yeah, the arguments or .petscrc are really great for experimenting!

Best,
Florian

> -- 
> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any
> results to which their experiments lead.
> -- Norbert Wiener