April 16th, 2009
· Lapack dependecy removed through native jacobi algorithm implementation.
· Boundary markers can now be added (and read) for meshes that are created by passing a surfacemodel_c to the meshgenerator.
· Cubic interpolation added
· generation of stiffness matrices for UP formulation of nonlinear elasticity is now possible. (untested!)
· Fixed doc for recovery_n, compound_n
· corin split matrix available again.
· sign of second order edge nodes have flipped.
· SUPG solver is integrated into Animath (with possibilty for non-zero pressure boundaries).
· Characteristc-Galerkin solver is integrated.
· Fixed treatment of shearing components in geometric stiffness matrix generator.
· Fixed a critical bug in mixed formulation stress recovery.
· Made destructors virtual in classes about which the compiler complained.
· Fixed embarrasing typo: neo-Hookian -> neo-Hookean
January 23rd, 2009
· SUPG weighting for upwind stabilization is now supported, and a corresponding solver can be found among the example applications.
· Another great methodolical enhancement is the addition of matrix generators for non-linear (including geometric non-linearity) structural analysis.
· Matrix and load vector generation have been totally redesigned for this release, the equation_c class and its descendants have been replaced with one class for each type of matrix/load.
· Many bugs have also been addressed, such as installation to global directories and outdated or incomplete documentation.
November 20th, 2008
· Much time went into the development of mesh manipulation code, i.e. edge collapse and vertex clustering routines (see the surfacemodel_c interface).
· Parallel solving of vector-valued problems that have decoupled components is now made easy by means of threads and employed in a new fluid solver.
· Some quite significant enhancements can also be found in the visualization code.
September 12th, 2008
· Most significantly, this version adds traction and stress recovery routines and second order surfaces with curvilinear coordinates.
· The writing of FSI-problem solvers is (again) conveniently possible.