syfi team mailing list archive

Thread
Date

[HG syfi] Bugfix in _manually_ differentiated form, the automatic Jacobi was the right one!

To: syfi-dev@xxxxxxxxxx
From: syfi@xxxxxxxxxx (SyFi)
Date: Tue, 29 Apr 2008 23:27:09 +0200 (CEST)
Delivered-to: syfi-dev@xxxxxxxxxx
Reply-to: Discussion of SyFi Development <syfi-dev@xxxxxxxxxx>

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

changeset:   59:24b7151c81d3fae5fe795d6fe5784cff035d538a
tag:         tip
user:        "Martin Sandve AlnÃ¦s <martinal@xxxxxxxxx>"
date:        Tue Apr 29 23:27:05 2008 +0200
files:       examples/solvers/u_cubed/demo.py
description:
Bugfix in _manually_ differentiated form, the automatic Jacobi was the right one!

Also tested a few choices of relaxation parameter,
and omega = 1.0 (no relaxation) easily beats omega != 1.0
in this case.


changeset:   58:61c2d88822c9cb50fa9093ca19b146485e8a2805
user:        "Martin Sandve AlnÃ¦s <martinal@xxxxxxxxx>"
date:        Tue Apr 29 23:03:11 2008 +0200
files:       examples/solvers/linear_elasticity/demo.py examples/solvers/u_cubed/FIXME examples/solvers/u_cubed/demo.py
description:
Implemented demo with Newtons method in Python,
solving the simple nonlinear equation u^3=f.

Discovered a likely bug in the automatic Jacobi
computation, the constants in the generated code differ
from the manually defined form in this case.


changeset:   57:e74c9eb9bda5a5db4830d24f8ed05d9fc2bda921
user:        "Martin Sandve AlnÃ¦s <martinal@xxxxxxxxx>"
date:        Tue Apr 29 15:08:16 2008 +0200
files:       examples/solvers/README examples/solvers/convection_diffusion/FIXME examples/solvers/reaction_diffusion/FIXME examples/solvers/reaction_diffusion/demo.py examples/solvers/u_cubed/FIXME examples/solvers/u_cubed/demo.py
description:
Added placeholders for nonlinear demos. Not implemented.


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