- Jan 16, 2015
-
-
Dave Moxey authored
-
Dave Moxey authored
-
- Jan 02, 2015
-
-
Chris Cantwell authored
-
- Jan 01, 2015
-
-
Chris Cantwell authored
-
Chris Cantwell authored
-
- Dec 29, 2014
-
-
Chris Cantwell authored
Various MeshConvert additions This MR adds various features and corrects quite a few bugs present in `MeshConvert`: *Features*: - Ability to split hexahedra in the boundary layer refinement module. - New module to add curvature defined by a scalar function z=f(x,y). - Improved Gmsh support for high-order elements. - A small test suite of relevant examples which we can build on later. *Fixes*: - Many fixes relating to face-interior curvature. - All `MeshConvert` elements are now orientated exactly in the same manner as their Nektar++ counterparts. (This will allow for some speed improvements in a future branch.) - Lots of small bugs in `InputNek` relating to high-order curvature. See merge request !401
-
Dave Moxey authored
-
- Dec 28, 2014
-
-
Chris Cantwell authored
-
- Dec 26, 2014
-
-
Dave Moxey authored
Fix/get exp index This MR fixes a couple of bugs in the GetExpIndex function for identifying the expansion in an ExpList containing a specific point. See merge request !417
-
Dave Moxey authored
-
- Dec 23, 2014
-
-
Dave Moxey authored
-
Dave Moxey authored
-
Dave Moxey authored
-
- Dec 22, 2014
-
-
Dave Moxey authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Dave Moxey authored
-
Dave Moxey authored
Feature/APE-advection This MR ports the advection operator in the APE class from EquationSystem::WeakDGAdvection to AdvectionWeakDG and removes the continuous galerkin code. As far as i can see, none of the existing advection classes would work for Continuous+APE. It also never had boundary conditions implemented, so i doubt anybody ever used it. Therefore i think removing it is our best choice here. I tested this branch on my local machine and on buildbot. Everything looks fine. See merge request !409
-
Dave Moxey authored
-
Dave Moxey authored
-
Dave Moxey authored
Feature/cardiac benchmark This MR adds features for completing the cardiac electrophysiology benchmark problems and support to handle using a pre-partitioned mesh to avoid excessive memory requirements on the root node. Specifically, it adds the following features: - Allow the mesh to be partitioned and then the solver exits (new option --part-only). - Support in the mesh partitioner to split mesh into a specific number of partitions rather than being fixed at the size of the row communicator (i.e. a serial process can now be used to pre-partition the mesh). - Allow a simulation to be started from the pre-partitioned *_xml file. - Added a ThresholdMin filter to complement the ThresholdMax filter. - Added a Benchmark filter in the Cardiac solver for outputting the activation and repolarisation times for comparing with established benchmark problems. See merge request !411
-
Dave Moxey authored
Fix/get face phys Fix to use correct memory when interpolating faces in tets, prisms and pyramids. See merge request !393
-
- Dec 21, 2014
-
-
Dave Moxey authored
-
-
- Dec 20, 2014
-
-
Chris Cantwell authored
-
Chris Cantwell authored
Fix CMake 3.0.2 warning This MR fixes a CMP0022 policy warning that appears in CMake 3.0.2, arising from a difference in the `INTERFACE_LINK_LIBRARIES` and `LINK_INTERFACE_LIBRARIES_DEBUG` properties that happens when debug/optimized versions of a library are specified in `TARGET_LINK_LIBRARIES`. See merge request !415
-
Chris Cantwell authored
Add block preconditioner for HDG This MR adds a block preconditioner for the HDG system, based on the inverse of each HDG face block, which reduces iteration counts significantly for hybrid 3D elements in comparison to the usual Jacobi preconditioner. 2/3D and parallel execution is supported. Documentation has also been added to the developer guide, based on what is presently in the Nektar++ wiki. Note that there are no tests as part of this branch, since there is no real testing framework for the preconditioners at all. Hopefully this will be added later! See merge request !397
-
Chris Cantwell authored
Another fix for ICC This MR follows on from !414 to fix the same issue in the VWI solver. See merge request !416
-
- Dec 19, 2014
-
-
Dave Moxey authored
-
Dave Moxey authored
-
- Dec 18, 2014
-
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
Chris Cantwell authored
-
- Dec 17, 2014
-
-
Dave Moxey authored
-
Dave Moxey authored
-