branches with status:
Name Status Last Modified Last Commit
lp://staging/~cwilson/spud/lines_attribute_checkpoint_fix (Has a merge proposal) 1 Development 2015-11-21 04:55:44 UTC
538. Fix string_type elements to have a li...

Author: Cian Wilson
Revision Date: 2015-11-21 04:52:21 UTC

Fix string_type elements to have a lines attribute.

This fixes checkpoints so that they aren't missing this attribute when first opened in diamond.

lp://staging/~cwilson/fluidity/testschema bug 1 Development 2012-03-01 19:30:02 UTC
3945. A first attempt at fixing the messed ...

Author: Cian Wilson
Revision Date: 2012-03-01 19:27:57 UTC

A first attempt at fixing the messed up tests schema. Update script still completely broken due to the ridiculous unpredictable way in which libspud deals with attributes.

lp://staging/~cwilson/fluidity/highra 1 Development 2011-10-28 02:57:27 UTC
3437. Merging in changes from trunk

Author: Cian Wilson
Revision Date: 2011-10-28 02:57:27 UTC

Merging in changes from trunk

lp://staging/~cwilson/fluidity/cv_subcells 1 Development 2011-04-19 00:03:23 UTC
3416. Prototype for control volume subcells...

Author: Cian Wilson
Revision Date: 2011-04-17 06:31:29 UTC

Prototype for control volume subcells. Only 2d triangles set up so far. This allows a number of things, principally galerkin projections on the control volume dual mesh, which are bounded and conservative without the need for a bounding algorithm. It also allows the visualisation of control volume fields on the meshes they were actually solved on. CV subcells can either be simplicial or cube. The latter is cheaper to supermesh (half the elements in 2d) but required fixes for some simplicial assumptions in conservative interpolation (this only works for p0 quads as the local coordinate calculations required for higher order elements have not been generalised - luckily p0 quads are enough for cv galerkin projection). Also the intersector assumes a particular orientation of quads in mesh A which must be enforced before generating the intersection map. Any form of projection (simplicial or cube) currently requires a continuous coordinate mesh. To generate this the cv sub cell mesh depends on an analogy with the p2bubble mesh, which is a hack. This does not generalise beyond 2d simplicial elements - hence the current restriction to triangles. Generalising this will require significant thought about global numbering :( Two tools have also been added... one for visualising triangle meshes and the other for visualising fields in a vtu on the cv sub cell mesh. These do not require a continuous mesh so could be generalised more easily. Tests and unittests have been added. Also the disabled quad galerkin projection test has been revived as a p0 quad projection test as this now works.

14 of 4 results