1. 16 Sep, 2015 3 commits
  2. 15 Sep, 2015 3 commits
  3. 14 Sep, 2015 1 commit
  4. 13 Sep, 2015 1 commit
  5. 11 Sep, 2015 1 commit
  6. 09 Sep, 2015 1 commit
  7. 06 Sep, 2015 5 commits
  8. 05 Sep, 2015 3 commits
  9. 01 Sep, 2015 4 commits
    • Chris Cantwell's avatar
      Merge branch 'feature/backtrace' into 'master' · 37671482
      Chris Cantwell authored
      this appends a nice backtraceto NekError to make it more obvious where the exception was thrown. Buildbot was successful. It looks like this:
      $ APESolver-g APE_2DChannel_WeakDG_MODIFIED.xml
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g() [0x482d7d]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g() [0x48346f]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g(_ZN6Nektar5ArrayINS_4OneDENS0_IS1_dEEEixEj+0x124) [0x48bc92]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g(_ZN6Nektar3APE12v_InitObjectEv+0x67d) [0x483d79]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g(_ZN6Nektar11SolverUtils14EquationSystem10InitObjectEv+0x23) [0x48a595]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g(_ZN6Nektar3APE6createERKN5boost10shared_ptrINS_12LibUtilities13SessionReaderEEE+0x5a) [0x48a77a]
      /home/lackhove/code/_build/nektar/library/SolverUtils/libSolverUtils-g.so.4.1.0(_ZN6Nektar12LibUtilities10NekFactoryISsNS_11SolverUtils14EquationSystemERKN5boost10shared_ptrINS0_13SessionReaderEEENS0_4noneESA_SA_SA_E14CreateInstanceESsS9_+0xe2) [0x7f81c9935c46]
      /home/lackhove/code/_build/nektar/library/SolverUtils/libSolverUtils-g.so.4.1.0(_ZN6Nektar11SolverUtils6Driver12v_InitObjectERSo+0x536) [0x7f81c9931428]
      /home/lackhove/code/_build/nektar/library/SolverUtils/libSolverUtils-g.so.4.1.0(_ZN6Nektar11SolverUtils14DriverStandard12v_InitObjectERSo+0x23) [0x7f81c9948acb]
      /home/lackhove/code/_build/nektar/library/SolverUtils/libSolverUtils-g.so.4.1.0(_ZN6Nektar11SolverUtils6Driver10InitObjectERSo+0x2e) [0x7f81c9946a64]
      /home/lackhove/code/_build/nektar/library/SolverUtils/libSolverUtils-g.so.4.1.0(_ZN6Nektar11SolverUtils14DriverStandard6createERKN5boost10shared_ptrINS_12LibUtilities13SessionReaderEEE+0x64) [0x7f81c994aba2]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g(_ZN6Nektar12LibUtilities10NekFactoryISsNS_11SolverUtils6DriverERKN5boost10shared_ptrINS0_13SessionReaderEEENS0_4noneESA_SA_SA_E14CreateInstanceESsS9_+0xe2) [0x479e24]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g(main+0x14a) [0x470582]
      /lib64/libc.so.6(__libc_start_main+0xf5) [0x7f81c2beab05]
      /home/lackhove/code/_build/nektar/solvers/APESolver/APESolver-g() [0x46f779]
      Fatal   : Level 1 assertion violation
      Where   : /home/lackhove/code/nektar/library/LibUtilities/BasicUtils/SharedArray.hpp[552]
      Message : Element 4 requested in an array of size 4
      See merge request !495
    • Dave Moxey's avatar
      Add some documentation · 2a909ecf
      Dave Moxey authored
    • Dave Moxey's avatar
      Fix bug with boundary to element/trace IDs · 6dbb7fa9
      Dave Moxey authored
    • Dave Moxey's avatar
  10. 31 Aug, 2015 1 commit
  11. 29 Aug, 2015 1 commit
  12. 27 Aug, 2015 1 commit
  13. 26 Aug, 2015 2 commits
  14. 25 Aug, 2015 5 commits
    • Chris Cantwell's avatar
      Merge branch 'fix/Interpolate3D' into 'master' · f0c44052
      Chris Cantwell authored
      PtsField: Use nektars geom tolerance for finding exact matches during shepard in…
      Use nektars geom tolerance for finding exact matches during shepard interpolation and check weigts for NaN after computation
      This fixes a problem where the weights become inf if the points are too close.
      Buildbot is all green
      See merge request !499
    • Chris Cantwell's avatar
    • Kilian Lackhove's avatar
    • Spencer Sherwin's avatar
      Added Regression tests. · fe7791a5
      Spencer Sherwin authored
    • Chris Cantwell's avatar
      Merge branch 'feature/cadsystem' into 'master' · aa829383
      Chris Cantwell authored
      This brings in three classes that act as a wrapper for opencascade to allow the loading and querying of cad information form STEP and IGES files. The CADSystem class loads and process the cad file before creating two maps of the classes CADCurve and CADSurf which contain the individual opencascade entities from the CAD file and querying functions.
      Because opencascade is such a huge set of libraries with many active versions, all of which can be compiled in numerous different ways. Cmake has been set up to download known working precompiled versions of opencascade dependant on the operating system, it then unpacks the libraries and headers into TPDIST to ensure they are included in the install step.
      See merge request !486
  15. 24 Aug, 2015 2 commits
    • Chris Cantwell's avatar
      Merge branch 'fix/XxtMultiLevel' into 'master' · b7d6c4b7
      Chris Cantwell authored
      Fix/xxt multi level
      This MR fixes the XxtMultiLevelStaticCond solver.
      See merge request !502
    • Chris Cantwell's avatar
      Merge branch 'fix/block-periodic' into 'master' · 75c08303
      Chris Cantwell authored
      Fix block preconditioner to work with periodic boundary conditions
      This MR introduces a streamlined version of the block preconditioner that works for problems with periodic boundary conditions in 2D and 3D. It also adds a new preconditioner metric for the test suite that allows for an integer tolerance defined in the `.tst` file, and fixes a potential bug with periodic edges in parallel inside `AssemblyMapCG`.
      See merge request !494
  16. 22 Aug, 2015 3 commits
  17. 21 Aug, 2015 1 commit
    • Spencer Sherwin's avatar
      Update Arpack Driver with the CoupledLinearNS solver. This now has a consitent... · a98b0d6b
      Spencer Sherwin authored
      Update Arpack Driver with the CoupledLinearNS solver. This now has a consitent format for outputing eigenvalues recognising we are using the Negative of the Stokes operator in these iterations. Also fixed a bug on the output of the eigenvalues so they now appear reasonable. Finally changed the Arpack iteration to be the more generalised inner product with respect to the mass matrix which seems to work better in Arpack independent of whether a shift is used. All of this has required the additional virtual function v_NegatedOp in EquationsSystem
  18. 20 Aug, 2015 2 commits