ufl team mailing list archive

Thread
Date

[HG UFL] Some optimizations after profiling.

To: ufl-dev@xxxxxxxxxx
From: ufl@xxxxxxxxxx (UFL)
Date: Fri, 20 Feb 2009 16:48:11 +0100 (CET)
Delivered-to: ufl-dev@xxxxxxxxxx

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

changeset:   610:23ea136875f21b106f6f0d4532395f8c4bf5051d
tag:         tip
user:        "Martin Sandve AlnÃ¦s <martinal@xxxxxxxxx>"
date:        Fri Feb 20 16:48:58 2009 +0100
files:       sandbox/test_hyperelasticity.py ufl/algorithms/traversal.py ufl/indexing.py ufl/tensors.py
description:
Some optimizations after profiling.
I think there's something stupid in the AD implementation,
3D hyperelasticity is a bit slow, which might mean there's
some algorithm parts with non-optimal order.


changeset:   609:6f3e9ef052811fd52f3c00016661c1e3d5f31248
user:        "Martin Sandve AlnÃ¦s <martinal@xxxxxxxxx>"
date:        Fri Feb 20 16:27:35 2009 +0100
files:       sandbox/clean.sh sandbox/dependencies.py sandbox/rename.py sandbox/summarizetodos.py sandbox/test.py sandbox/testclasses.py sandbox/testproduct.py ufl/common.py
description:
Added all files in sandbox, to be on the safe side. probably some garbage here.


changeset:   608:05b3acaab9a609d7a9033d8eee3dad65bc7e3f38
user:        "Martin Sandve AlnÃ¦s <martinal@xxxxxxxxx>"
date:        Fri Feb 20 16:04:08 2009 +0100
files:       ufl/basisfunction.py ufl/common.py ufl/function.py
description:
Bizzarre bugfix in Counted class:
Function and Constant sometimes used
the same counter and sometimes not...

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