[petsc-dev] Do we really need MAT_IGNORE_MATRIX?

Barry Smith bsmith at mcs.anl.gov
Thu Mar 18 16:17:53 CDT 2010


Con: Then what should one pass for MatReuse? Neither initial_matrix or  
reuse_matrix makes sense as the flag for the null one passed. Could  
just say it doesn't matter what flag is used.

Pro: With this addition, all the other routines that take MatReuse are  
broken, since they generally just do one check on the reuse flag and  
then else to the other case. They do not handle the ignore case well.

    I kind of like the new value, but would like to see the other  
routines fixed to generate an error up front if they are called with it.


On Mar 17, 2010, at 11:10 PM, Lisandro Dalcin wrote:

> I've been very busy with other stuff, so I've not followed PETSc
> development too closely as I should...
> IIUC, the only purpose of MAT_IGNORE_MATRIX is for getting only one of
> mat or pmat...  Would'nt be enough to pass a NULL pointer for newmat
> or newpmat?
> -- 
> Lisandro Dalcin
> ---------------
> Centro Internacional de Métodos Computacionales en Ingeniería (CIMEC)
> Instituto de Desarrollo Tecnológico para la Industria Química (INTEC)
> Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET)
> PTLC - Güemes 3450, (3000) Santa Fe, Argentina
> Tel/Fax: +54-(0)342-451.1594

More information about the petsc-dev mailing list