dolfin team mailing list archive

[Johan Hake <hake@xxxxxxxxx>]

On Thursday 05 March 2009 11:53:38 DOLFIN wrote:
> One or more new changesets pushed to the primary dolfin repository.
> A short summary of the last three changesets is included below.
>
> changeset:   5811:14d139b2e5fe24d018c37b0bf81dcc8594a38512
> tag:         tip
> parent:      5810:2bc3cceb38b6cacab1da81b2187e4fb2344df393
> parent:      5809:3d185b5c627237bbacb972fed1ea24342c77e627
> user:        "Garth N. Wells <gnw20@xxxxxxxxx>"
> date:        Thu Mar 05 10:53:26 2009 +0000
> files:
> description:
> merge.
>
>
> changeset:   5810:2bc3cceb38b6cacab1da81b2187e4fb2344df393
> parent:      5807:ef9b4936b044cb7d025fe2796840059fb3773554
> user:        "Garth N. Wells <gnw20@xxxxxxxxx>"
> date:        Thu Mar 05 10:53:10 2009 +0000
> files:       demo/mesh/refinement/cpp/main.cpp dolfin/la/EpetraMatrix.cpp
> dolfin/la/EpetraMatrix.h dolfin/la/GenericMatrix.h dolfin/la/MTL4Matrix.cpp
> dolfin/la/MTL4Matrix.h dolfin/la/Matrix.h dolfin/la/PETScMatrix.cpp
> dolfin/la/PETScMatrix.h dolfin/la/STLMatrix.h dolfin/la/uBLASMatrix.h
> description:
> Add 'bool same_nonzero_pattern' to GenericMatrix::axpy(...).

Nice!

But I am not sure this is a satisfactory solution. It is not possible to set 
the same_nonzero_pattern flag when using the operator={+,-}. This is shown in 
the failing la unit test for the Epetra backend. The operators in PETSc will 
also be unnecessary slow.

Is it possible to have a global parameter same_nonzero_pattern, or should we 
default the axpy call in these operators with same_nonzero_pattern=true, or a 
third more dramatical alternative, removing the operators and only using 
axpy?

Johan

>
> changeset:   5809:3d185b5c627237bbacb972fed1ea24342c77e627
> parent:      5808:6270c8454a8cb2e1d3e5bc40a505a7d4376a38ae
> parent:      5807:ef9b4936b044cb7d025fe2796840059fb3773554
> user:        "Ola Skavhaug <skavhaug@xxxxxxxxx>"
> date:        Thu Mar 05 10:54:13 2009 +0100
> files:       dolfin/io/XMLHandler.h
> description:
> merge
>
> ----------------------------------------------------------------------
> For more details, visit http://www.fenics.org/hg/dolfin
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev