1. 11 Oct, 2016 1 commit
  2. 24 May, 2016 1 commit
  3. 25 Apr, 2016 2 commits
  4. 14 Sep, 2015 3 commits
  5. 20 Jul, 2015 1 commit
  6. 04 May, 2015 1 commit
  7. 29 Sep, 2014 1 commit
  8. 11 Sep, 2013 1 commit
  9. 03 Sep, 2013 1 commit
  10. 14 Jun, 2013 1 commit
    • Tim Tautges's avatar
      Big batch of smallish changes working toward performance analysis of tree... · 1480624e
      Tim Tautges authored
      Big batch of smallish changes working toward performance analysis of tree stuff.  Make check runs correctly.
      Fixed includes of FileOptions, indicated by an out-of-source build.
      Added a cpu timer class (lifted from WriteHDF5), and added that to call stats in the trees.
      Fixed problem creating multiple sets at a time, have to duplicate flags into an array instead of using just one.
      Added Tree::parse_options as a required virtual function.
      Added optional output() method to TreeStats to output all stats on a single line.
      Added testing Makefile.am magic for test/perf/point_location subdirectory, and several new performance tests for tree stuff.
      Added several options to spatial_locator_test to allow various size meshes and other options.
      Fixed several places in trees (or put in asserts) that were evaluating tree root before setting it
      In reverse evaluation, if not converged but are outside at time of test and inside flag returned from function, return success
      Fixed a few places where connectivity requested but no dummy array input (so didn't work for scd mesh)
      In SpatialLocator point search, pass in tolerance to element reverse evaluation
      Fixed dumb makefile.am misspelling
      sploc_searching_perf: add support for element evaluation and for tolerance
  11. 16 May, 2013 1 commit
    • Iulian Grindeanu's avatar
      replace free() calls from fortran with iMesh_freeMemory() calls · df79abdb
      Iulian Grindeanu authored
      they will call underneath free() from c;
      It seems that ibm fortran (xlf) does not like the mix between malloc in c and
      free in fortran
      function is added as an iMesh extension, and it should be used from fortran imesh
      In iMesh C or C++ it is still fine  to just call free() directly.
  12. 05 Dec, 2012 1 commit
  13. 29 Jun, 2012 1 commit
  14. 27 Apr, 2012 1 commit
  15. 25 Apr, 2012 1 commit
  16. 18 Apr, 2012 1 commit
  17. 04 Apr, 2012 1 commit
  18. 02 Apr, 2012 1 commit
  19. 26 Mar, 2012 1 commit
  20. 05 Mar, 2012 1 commit
  21. 23 Feb, 2012 1 commit
    • Tim Tautges's avatar
      Mostly some imesh/ibase changes: · 1a555093
      Tim Tautges authored
      - moved some arguments around in connect_iterate and coords_iterate functions
      - moved iterator #define's to iBase_f.h.in from iMesh_f.h
      - implemented new imesh extension functions iMesh_connectIterate and iMesh_coordsIterate, which provide direct access to 
        memory used to store connectivity/coordinates, resp.
      - added an argument to get the # vertices/entity to connect_iterate
      - implemented new iMesh extension function iMesh_initEntArrIterRec, which is an iterator that recurses into contained
        entity sets; all other iterator-related functions operate as before, but with recursive access to sets (this also
        required changes to implementation in MBIter, but since MOAB's underlying functions already had a recursive flag,
        this just involved storing/sending the flag to the MOAB function
      - in vtkMOABReader, implemented more careful determination of vtk cell type, and support quadratic cell types in some
      Passes make check.
      git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@5410 6162379e-bd28-0410-9a7d-b7f4dcfcad3c
  22. 17 Feb, 2012 1 commit
  23. 02 Feb, 2012 1 commit
  24. 13 Jan, 2012 1 commit
    • Tim Tautges's avatar
      Mostly changes to remove compiler warnings when compiling with PGI. · 72a7515b
      Tim Tautges authored
      test/kd_tree_time.cpp: adding include, to enable compiling with pgi
      itaps/imesh/iMesh_MOAB.cpp: fixing CreateTagWithOptions, which wasn't passing default
        value correctly
      itaps/imesh/ScdMeshF77.F: had a few ';' characters that caused warnings under PGI
      itaps/imesh/Makefile.am, examples/itaps/Makefile.am: adding -I$(top_builddir)/itaps, which is 
        needed now that we are generating iBase_f.h
      src/io/ReadMCNP5.cpp, ReadNASTRAN.cpp: removing a few break statements that are never reached, to get rid
        of compile warnings
      src/io/ReadNCDF.cpp: adding some code to use a result, to remove a compiler warning
      src/io/WriteTemplate.cpp, src/io/ReadNC.cpp: similar
      src/SparseTag.cpp: removing a variable, to remove a compiler warning
      tools/dagmc/quads_to_tris.cpp: adding some code to remove compiler warning
      tools/dagmc/test_geom.cc: increasing tolerance value in comparison, otherwise sometimes
        this test fails on comparison to std::numeric_limits<double>::epsilon()
      Passes make check in serial and parallel.
      git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@5320 6162379e-bd28-0410-9a7d-b7f4dcfcad3c
  25. 22 Dec, 2011 1 commit
    • Tim Tautges's avatar
      Fixing bug in parallel reading, where if a set contains only other sets, and · 99639e7f
      Tim Tautges authored
      the contained sets are read locally (based on containing locally-represented
      entities), the containing set is also selected for reading.  This comes up
      with NEUMANN sets read from cub files, where those sets contain only the
      surface geom_topo sets.  Fixes problem where Nek wasn't running with READ_PART
      as the parallel read option.
      A few other assorted changes:
      - change to listing of set entities in Core::list_entities
      - better option processing in iMeshP
      - added a few more iMesh example codes
      Passes checks in serial and parallel.
      git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@5280 6162379e-bd28-0410-9a7d-b7f4dcfcad3c
  26. 08 Dec, 2011 1 commit
  27. 16 Nov, 2011 1 commit
  28. 14 Nov, 2011 1 commit
    • Tim Tautges's avatar
      test/Makefile.am: if parallel, include the src/parallel in places to look for includes · 00235ef4
      Tim Tautges authored
      itaps/imesh/iMesh_MOAB.cpp, iMesh_extensions.h: implement new extension function iMesh_createStructuredMesh,
       which creates a rectangle of structured mesh and, optionally, resolves shared vertices on the boundary
       and exchanges ghost elements.
      itaps/iBase_f.h: replace 'c' with '!' as a comment symbol, for comatability with F90 free-source form
      src/parallel/moab/ParallelComm.hpp: add comment for one of the arguments
      Passes make check, though I haven't written any tests for the iMesh extension, I'm with the customer that needs
      it, so I'll have to implement that later.
      git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@5212 6162379e-bd28-0410-9a7d-b7f4dcfcad3c
  29. 28 Sep, 2011 1 commit
  30. 11 Aug, 2011 1 commit
  31. 04 Aug, 2011 1 commit
    • Tim Tautges's avatar
      This commit adds iMesh access to tag_iterate, and the ability to create dense tags through iMesh. · cd9b0df4
      Tim Tautges authored
      See the itaps/imesh/iMesh_extensions.h function for details of these functions and how to use them.
      Three new iMesh extension functions are added: iMesh_createTagWithOptions, iMesh_stepIter, and iMesh_tagIterate.
      The first two were needed before the tagIterate function could be used.  Unit test and example codes are also added
      to test this.
      Besides changes for the above, this commit also has the following changes:
      iMesh_f.h: change iMesh_EntityIterator to iBase_EntityIterator, same for array-based version.
      iMesh-Defs.inc.in: add @PNETCDF_LIBS@ to IMESH_LIBS
      itaps/imesh/Makefile.am: added more -I locations, and new MOAB_iMesh_extensions_tests test program
      tools/vtkMOABReader/Makefile.am: added another library to LDADD
      configure.ac: replaced check for vtkRenderer.h with one for vtkRenderingInstantiator.h, that seems to be a 
        better header to look for; also added another default location to check for vtk; also added examples/itaps
        to the list of makefiles to create
      examples/SetsNTags.cpp: changed to namespace'd MOAB calls
      Passes make check in serial and parallel.
      git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@5082 6162379e-bd28-0410-9a7d-b7f4dcfcad3c
  32. 21 Jul, 2011 1 commit
  33. 24 Jun, 2011 1 commit
  34. 09 Jun, 2011 1 commit
  35. 08 Jun, 2011 2 commits
  36. 06 Jun, 2011 1 commit
    • 's avatar
      Remaining tag API changes: convert from sizes in bytes to sizes as number of · a39f15ff
      values where appropriate:
      o Mark Interface::tag_get_size as deprecated and add new functions:
        tag_get_length and tag_get_bytes, where the former returns the number
        of values of the tag data type (e.g. number of doubles for tag with
        MB_TYPE_DOUBLE), returning the number of bits for bit tags and the
        number of bytes for opaque tags.  The latter is the same as the old
        tag_get_size, *except* that it always returns 1 for bit tags.
      o Make the tag_get/set_data functions that work with arrays of pointers
        to tag values and arrays of lengths as deprecated, replacing them with
        analogous tag_get/set_by_ptr functions that work with lengths in terms
        of number of values rather than number of bytes.  Function rename will
        also help to avoid memory corruption when attempting to call the old
        functions w/out the optional length array and accidentally calling the
        non-pointer variants due to accidental cast to void*.
      o Change seldom used optional size argument to tag_get_default_value and
        tag_clear_data to be interpreted as number of values rather than number
        of bytes.
      o Update all MOAB code for these changes.
      git-svn-id: https://svn.mcs.anl.gov/repos/ITAPS/MOAB/trunk@4954 6162379e-bd28-0410-9a7d-b7f4dcfcad3c