--- /dev/null
+// Copyright (C) 2007-2023 CEA, EDF
+//
+// This library is free software; you can redistribute it and/or
+// modify it under the terms of the GNU Lesser General Public
+// License as published by the Free Software Foundation; either
+// version 2.1 of the License, or (at your option) any later version.
+//
+// This library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+// Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public
+// License along with this library; if not, write to the Free Software
+// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+//
+// See http://www.salome-platform.org/ or email : webmaster.salome@opencascade.com
+//
+
+#include "ByStringMPIProcessorGroup.hxx"
+
+#include <iostream>
+#include <set>
+#include <algorithm>
+#include "mpi.h"
+
+using namespace std;
+
+
+namespace MEDCoupling
+{
+ /*!
+ \class ByStringMPIProcessorGroup
+
+ The ByStringMPIProcessorGroup implements a derived version of MPIProcessorGroup.
+
+ Groups are formed from MPI ranks with the same simCodeTag. Two trivial cases:
+ - All simCodeTag are equal, then one group is formed from all mpi ranks in the communicator
+ - All simCodeTag are different, then n-MPI groups are formed.
+ */
+
+
+ /*! Build and return a map between different simCodeTag and the set of MPI ranks ids (based on the passed communicator) grouped by the same identifier.
+ \param interface CommInterface object giving access to the MPI communication layer
+ \param simCodeTag the string identifiying the tag for the group.
+ \param world_comm mpi communicator
+ \return Map relating unique simCodeTag and the group of MPI ranks ids belowing to that group
+ */
+ static std::map<std::string,std::set<int>> DefineSetIdByStringName( const CommInterface& interface, const std::string& simCodeTag, const MPI_Comm& world_comm )
+ {
+ int size_world;
+ int rank_world;
+ interface.commSize(world_comm,&size_world);
+ interface.commRank(world_comm,&rank_world);
+
+ std::map<std::string,std::set<int>> myRanksSet;
+
+ std::vector<int> displacement(size_world, 0 );
+ std::vector<int> words_size(size_world);
+
+ int stringSize = (int) simCodeTag.size();
+ interface.allGather( &stringSize, 1, MPI_INT, words_size.data(), 1, MPI_INT, world_comm );
+
+ for (size_t rank = 1; rank < words_size.size(); rank++)
+ displacement[ rank ] = words_size[ rank - 1 ] + displacement[ rank - 1 ];
+
+ char globalnames[displacement[size_world-1]];
+
+ interface.allGatherV( simCodeTag.c_str(), stringSize, MPI_CHAR, &globalnames,
+ words_size.data(), displacement.data(), MPI_CHAR, world_comm );
+
+ for (size_t rank = 0; rank < size_world; rank++)
+ {
+ std::string strByRank( &globalnames[displacement[ rank ]], words_size[ rank ] );
+ myRanksSet[ strByRank ].insert( (int)rank );
+ }
+ return myRanksSet;
+ }
+
+ /*!
+ * Creates a processor group that is based on all the
+ processors of MPI_COMM_WORLD .This routine must be called by all processors in MPI_COMM_WORLD.
+ \param interface CommInterface object giving access to the MPI
+ communication layer
+ */
+ ByStringMPIProcessorGroup::ByStringMPIProcessorGroup(const CommInterface& interface):
+ MPIProcessorGroup(interface)
+ {
+ }
+
+ /*! Creates a processor group based in the simCodeTag passed.
+
+ \param interface CommInterface object giving access to the MPI
+ communication layer
+ \param simCodeTag the string identifiying the tag for the group.
+ \param world_comm mpi communicator
+ */
+ ByStringMPIProcessorGroup::ByStringMPIProcessorGroup(const CommInterface& interface, const std::string& simCodeTag, const MPI_Comm& world_comm ):
+ MPIProcessorGroup(interface, DefineSetIdByStringName( interface, simCodeTag, world_comm ), simCodeTag, world_comm )
+ {
+ }
+
+ ByStringMPIProcessorGroup::ByStringMPIProcessorGroup(const ByStringMPIProcessorGroup& other):
+ MPIProcessorGroup(other)
+ {
+ }
+
+ ByStringMPIProcessorGroup::~ByStringMPIProcessorGroup()
+ {
+ }
+
+ ByStringMPIProcessorGroup *ByStringMPIProcessorGroup::deepCopy() const
+ {
+ return new ByStringMPIProcessorGroup(*this);
+ }
+
+}
