April 2016 - moab-dev - mailman3.cels.anl.gov

a bug in Coupler::locate_points
by Huayi Wei 17 Apr '17

17 Apr '17

Hi, Moab Dev, In the following member function, I think it is a bug for line ` tl = new TupleList(3, 0, 0, 0, num_points);`. Here create new TupleList which address will cover the address tl passed by user, and then user can not accsess it. I think it should be ' tl->initialize(3, 0, 0, 0, num_points)`, ```` ErrorCode Coupler::locate_points(double *xyz, unsigned int num_points, double rel_eps, double abs_eps, TupleList *tl, bool store_local) { ....... // Copy into tl if passed in and storing locally if (tl && store_local) { tl = new TupleList(3, 0, 0, 0, num_points); tl->enableWriteAccess(); memcpy(tl->vi_wr, tl_tmp->vi_rd, 3 * tl_tmp->get_n() * sizeof(int)); tl->set_n(tl_tmp->get_n()); tl->disableWriteAccess(); } .... } ``` Best Huayi

4 6

Memory Leak in moab::Core::Initialize( )
by Dario Pezzoli 22 Apr '16

22 Apr '16

Dear all, I get a memory leak issue when it comes to free the memory allocated by the constructor moab::Core::Core( ). Have someone already encountered this problem? Many thanks in advance. Yours, Dario Here is the Valgring report: ==20018== at 0xDDA469B: opal_os_dirpath_create (in /usr/lib/openmpi/lib/libmpi.so.1.0.8) ==20018== by 0xDD3F4EC: orte_session_dir (in /usr/lib/openmpi/lib/libmpi.so.1.0.8) ==20018== by 0xDD529E9: orte_ess_base_app_setup (in /usr/lib/openmpi/lib/libmpi.so.1.0.8) ==20018== by 0x2A07EBA5: ??? (in /usr/lib/openmpi/lib/openmpi/mca_ess_singleton.so) ==20018== by 0xDD32842: orte_init (in /usr/lib/openmpi/lib/libmpi.so.1.0.8) ==20018== by 0xDCF2023: ompi_mpi_init (in /usr/lib/openmpi/lib/libmpi.so.1.0.8) ==20018== by 0xDD0965F: PMPI_Init (in /usr/lib/openmpi/lib/libmpi.so.1.0.8) ==20018== by 0x5D6F060: moab::Core::initialize() (in /opt/moab/lib/x86_64-linux-gnu/libMOAB.so.4.9.0) ==20018== by 0x5D6F11C: moab::Core::Core() (in /opt/moab/lib/x86_64-linux-gnu/libMOAB.so.4.9.0) ==20018== by 0x45BFBD: mesh::MoabMesh::initializeMoabInterface()

2 1

periodic boundaries
by Alexander Nikolas Breuer 20 Apr '16

20 Apr '16

Dear MOAB-team, I am trying to read a triangular mesh with periodic boundaries created by gmsh into MOAB (attached). However I don’t see this when going through MOAB’s adjacencies. Is there a way to make the mesh reader aware of periodic boundaries? Thanks, Alex -- Alexander Breuer University of California, San Diego San Diego Supercomputer Center 9500 Gilman Drive MC 0505 La Jolla, California 92093-0505, USA http://hpgeoc.sdsc.edu +1 (858) 822-5692

3 9

MOAB & periodicity
by Xavier Garnaud 12 Apr '16

12 Apr '16

Dear all, I am thinking about using MOAB as the basis for a cell-centered (unstructured) finite volume solver. An important feature for me is to be able to handle periodicity (translational & rotational). As far as I understand, this is not directly possible with MOAB but I am wondering how difficult it would be. What I would like to do is have "periodic ghost" cells that I would handle as the standard ghost cells (except I would possibly apply a rotation to the vectors after the data is exchanged for rotational periodicity) How difficult do you think it would be to implement it? Do you have any example that implement similar operations? Best regards Xavier

4 3

MOAB v4.9.1 release (SIGMA v1.2.1)
by Vijay S. Mahadevan 10 Apr '16

10 Apr '16

Dear users, The consolidated release of SIGMA v1.2.1 is available now and contains the MOAB 4.9.1 release candidate. The release tarball can be downloaded at: ftp://ftp.mcs.anl.gov/pub/fathom/moab-4.9.1.tar.gz. Apart from several important fixes in the codebase, some notable changes in MOAB v4.9.1 are: * MOAB now supports auto-download and configuration/installation of several key dependencies (HDF5, NetCDF and Metis/ParMetis) out of the box. Users can specify the following configuration options to enable this feature. We plan to add more dependency support in future releases. > ./configure --download-hdf5 --download-netcdf --download-metis * A configuration suggestion script (suggest_configuration.sh) is now part of MOAB repository. This script simplifies options to be used for new users by providing suggestions on standard LCF machines and to some extent, on arbitrary machines depending on the environment. * Scalable parallel uniform mesh refinement with parent-to-child hierarchy queries that can support geometric multigrid solvers. * Fixes for Core::connect_iterate on non-congruently allocated cells of same type. * Improve augmenting ghost entities with set information (remove parallel and memory bottlenecks) * Enhanced support for dealing with polygons and polyhedra with added functionality for VTK format (Exodus support soon) Please refer to the following link to find all changes in the SIGMA components and more specifically,. information about the fixes and enhancements in MOAB as part of this release. http://sigma.mcs.anl.gov/news/sigma-v1-2-1 Let us know if you have any questions or encounter any issues during your upgrade. Best, The SIGMA team

1 0

file formats
by Nico Schlömer 08 Apr '16

08 Apr '16

Hi everyone, I'll start tackling a time-series problem rather soon and I'm wondering about file formats here. Ideally, I'd have one time series per file (particularly since the mesh isn't changing over time, perhaps that could be made use of), but I'm not sure if MOAB's custom HDF5 supports that yet. Thinking of which, I faintly remember there had been talk about switching to a better-supported format in the future (xdmf and exodus came to mind). Is this still a thing? Cheers, Nico

5 10

subdomains, sparse tags
by Nico Schlömer 07 Apr '16

07 Apr '16

Hi everyone, I'm just reading up on tags, all starting from the problem that I have mesh on which there are multiple subdomains which may or may not overlap, may or may not fill the the entire mesh. What I will need to do is to set those tags on the mesh and then to iterate over all entities with that tag (probably multiple times). Do I see correctly that sparse tags [1] are the way to go then? Cheers, Nico [1] https://www.mcs.anl.gov/~fathom/moab-docs/html/classmoab_1_1SparseTag.html

2 3

Re: [MOAB-dev] Resolving shared entities on prisms
by Grindeanu, Iulian R. 05 Apr '16

05 Apr '16

hmmm, I didn't realize you have higher order nodes for triangles and some edges Are all quads linear? Yes, I see multiple problems, and they are related to higher order nodes. the prisms are all linear, and in read_part, they are the primary entities, so the higher order nodes are ignored. (they are not part of the partition). So then the triangles are not read into the part, because they have nodes that are not in the local partition. (in read_part method) bcast_delete is reading all mesh from file, and then deleting the "non local" elements; Still need to understand what is going on Iulian ________________________________________ From: Lukasz Kaczmarczyk [Lukasz.Kaczmarczyk(a)glasgow.ac.uk] Sent: Tuesday, April 05, 2016 11:13 AM To: Grindeanu, Iulian R. Subject: Re: Resolving shared entities on prisms Hello, File which I send you before is with mid-nodes on triangles. This is not working with PARALLEL=READ_PART. I created another file, with nodes only on corers. See attached file and this works with PARALLEL=READ_PART, however it is not working with BCAST_DELETE. It could be not one, but two separate problems. One with resolving ents when BCAST_DELETE or READ_DELETE is used, another with mid-modes. Kind regards, Lukasz > On 5 Apr 2016, at 15:44, Grindeanu, Iulian R. <iulian(a)mcs.anl.gov> wrote: > > Hi Lukasz > If you load using PARALLEL=READ_PART, instead of BCAST_DELETE, it is fine > > It is a bug somewhere, I am a little puzzled why it would show up only for BCAST_DELETE > > Still looking ... > Thanks, > Iulian > ________________________________________ > From: moab-dev-bounces(a)mcs.anl.gov [moab-dev-bounces(a)mcs.anl.gov] on behalf of Grindeanu, Iulian R. [iulian(a)mcs.anl.gov] > Sent: Tuesday, April 05, 2016 9:20 AM > To: Lukasz Kaczmarczyk; moab-dev(a)mcs.anl.gov > Subject: Re: [MOAB-dev] Resolving shared entities on prisms > > take that back; it is OK for hexas; so it may be related to prisms only > ________________________________________ > From: moab-dev-bounces(a)mcs.anl.gov [moab-dev-bounces(a)mcs.anl.gov] on behalf of Grindeanu, Iulian R. [iulian(a)mcs.anl.gov] > Sent: Tuesday, April 05, 2016 9:13 AM > To: Lukasz Kaczmarczyk; moab-dev(a)mcs.anl.gov > Subject: Re: [MOAB-dev] Resolving shared entities on prisms > > Hi Lukasz, > Thank you for your example; > I think it is related to the way we look at the skin. > It would show the same problem for other elements (hexas, with more than one layer) > > So, when we collect potential shared entities at the interface, for edges we should just collect by adjacencies, from the skin formed by 2d elements; we look at the 1d skin, which I think is wrong > > I will try to fix it soon > > Iulian > > > > ________________________________________ > From: moab-dev-bounces(a)mcs.anl.gov [moab-dev-bounces(a)mcs.anl.gov] on behalf of Lukasz Kaczmarczyk [Lukasz.Kaczmarczyk(a)glasgow.ac.uk] > Sent: Tuesday, April 05, 2016 6:51 AM > To: moab-dev(a)mcs.anl.gov > Subject: [MOAB-dev] Resolving shared entities on prisms > > Hello, > > I have created muli-layered shell from prisms elements. For attached partitioned mesh (3 partitions) for I check output for shared own entities, however internal edges, on shared quads, between layers (internal edges) are not resolved as a shared. Everything else is ok. > > This is what I don’t understand, it could be error on my side, however I don;t see where. Could you pleas look at this and point me to potential error. It could be related to prisms? It works perfectly for single layer of prisms elements. > > PS. > This is part of work on solid shell elements, for interested, pleas look here, > https://www.dropbox.com/s/zml82aj027hfsbb/solid_shell.pptx?dl=0 > > > Kind regards, > Lukasz > > const char *option; > option = “PARALLEL=BCAST_DELETE;" > "PARALLEL_RESOLVE_SHARED_ENTS;" > "PARTITION=PARALLEL_PARTITION;"; > rval = moab.load_file(“tmp.h5m", 0, option); CHKERRQ_MOAB(rval); > > { > > Range shared_ents; > // Get entities shared with all other processors > rval = pcomm->get_shared_entities(-1, shared_ents); > Range owned_entities; > rval = pcomm->filter_pstatus(shared_ents, PSTATUS_NOT_OWNED, PSTATUS_NOT, -1, &owned_entities); > EntityHandle meshset; > rval = moab.create_meshset(MESHSET_SET,meshset); CHKERRQ_MOAB(rval); > rval = moab.add_entities(meshset,owned_entities); CHKERRQ_MOAB(rval); > > ostringstream o1; > o1 << "owned_" << pcomm->rank() << ".vtk"; > moab.write_file(o1.str().c_str(),"VTK","",&meshset,1); > > rval = moab.delete_entities(&meshset,1); CHKERRQ_MOAB(rval); > } > >

2 1

Resolving shared entities on prisms
by Lukasz Kaczmarczyk 05 Apr '16

05 Apr '16

Hello, I have created muli-layered shell from prisms elements. For attached partitioned mesh (3 partitions) for I check output for shared own entities, however internal edges, on shared quads, between layers (internal edges) are not resolved as a shared. Everything else is ok. This is what I don’t understand, it could be error on my side, however I don;t see where. Could you pleas look at this and point me to potential error. It could be related to prisms? It works perfectly for single layer of prisms elements. PS. This is part of work on solid shell elements, for interested, pleas look here, https://www.dropbox.com/s/zml82aj027hfsbb/solid_shell.pptx?dl=0 Kind regards, Lukasz const char *option; option = “PARALLEL=BCAST_DELETE;" "PARALLEL_RESOLVE_SHARED_ENTS;" "PARTITION=PARALLEL_PARTITION;"; rval = moab.load_file(“tmp.h5m", 0, option); CHKERRQ_MOAB(rval); { Range shared_ents; // Get entities shared with all other processors rval = pcomm->get_shared_entities(-1, shared_ents); Range owned_entities; rval = pcomm->filter_pstatus(shared_ents, PSTATUS_NOT_OWNED, PSTATUS_NOT, -1, &owned_entities); EntityHandle meshset; rval = moab.create_meshset(MESHSET_SET,meshset); CHKERRQ_MOAB(rval); rval = moab.add_entities(meshset,owned_entities); CHKERRQ_MOAB(rval); ostringstream o1; o1 << "owned_" << pcomm->rank() << ".vtk"; moab.write_file(o1.str().c_str(),"VTK","",&meshset,1); rval = moab.delete_entities(&meshset,1); CHKERRQ_MOAB(rval); }

2 4