← Back to team overview

dolfin team mailing list archive

[HG DOLFIN] Use 5 processes instead of 2 in system test for parallel assembly/solve. Works fine now.

 

One or more new changesets pushed to the primary dolfin repository.
A short summary of the last three changesets is included below.

changeset:   6996:8974976b83dd32de4395c25c364ed9737184417a
tag:         tip
user:        Anders Logg <logg@xxxxxxxxx>
date:        Thu Sep 10 09:45:23 2009 +0200
files:       test/system/parallel-assembly-solve/solver.py test/system/parallel-assembly-solve/test.py
description:
Use 5 processes instead of 2 in system test for parallel assembly/solve. Works fine now.
Running with 9 processes gives

olfin/fem/Assembler.cpp:202: static void dolfin::Assembler::assemble_exterior_facets(dolfin::GenericTensor&, const dolfin::Form&, dolfin::UFC&, const dolfin::MeshFunction<unsigned int>*, std::vector<double, std::allocator<double> >*): Assertion `cell_map' failed.

The mesh is probably too small for 9 processes (?) but it shouldn't fail with an assertion
on cell_map.


changeset:   6995:aef4a5fe0295075b695478698fc82d13717f2874
user:        Niclas Jansson <njansson@xxxxxxxxxx>
date:        Thu Sep 10 09:12:58 2009 +0200
files:       dolfin/fem/DofMap.cpp
description:
Fix boundary computation.


changeset:   6994:3d4a2430f794340aa0f545ef2416f8ece5deb0c0
user:        Anders Logg <logg@xxxxxxxxx>
date:        Wed Sep 09 23:03:00 2009 +0200
files:       dolfin/fem/DofMap.cpp dolfin/fem/DofMap.h
description:
Add temporary workaround tabulate_coordinates, works from Python with flattened array.

----------------------------------------------------------------------
For more details, visit http://www.fenics.org/hg/dolfin